{ "name": "BestForWhat", "schema_version": "1.0.0", "research_cutoff": "2026-09-28", "generated_at": "2026-09-28T00:06:04.405Z", "status": "two-run-synthesis", "source_policy": "competitor-only; all product claims remain unverified by independent testing", "hands_on_testing": false, "prompt_manifest": { "prompt_version": "1.0.0", "cutoff": "2026-09-28", "files": { "prompts/categories/email-marketing.md": "b796bffb6646f43a4a2addf605e114143383d86ca504265a0c76292a51369cc9", "prompts/categories/meeting-notes.md": "0b562aefb101d6f9da6613a8a8006bcb631f7ce61ca937ca8207b52bd3928d53", "prompts/categories/online-forms.md": "396002f47b15e1ead0b45f882cb9394ced0b64b2aa39ee75cbe805059a5404e5", "prompts/categories/project-management.md": "494fb81befd2f25f8f51abd3a206b4425c017eb4b7fa3822d918a28135a92ede", "prompts/categories/workflow-automation.md": "555d80817932b2d4ed21ba549a3f955dac45ffda04b63acfc8aee8dce09dbe3c", "prompts/integration-v1.0.md": "4e1674d27c947896132b295a60c36de98425bab5ce5070b75645bcd491d5b0ef", "prompts/research-v0.1.md": "5bb107045b848142c2ff3ba23210eb545c82677d74fd1e8f1cf7ad0b41bb70db", "prompts/research-v1.0.md": "198486731855cbe6c6d448f2442323740d44bb80de00f84dec48e7d072bc7615" }, "workflow-automation": { "input": "research/inputs/workflow-automation.md", "sha256": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723" }, "project-management": { "input": "research/inputs/project-management.md", "sha256": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d" }, "meeting-notes": { "input": "research/inputs/meeting-notes.md", "sha256": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c" }, "online-forms": { "input": "research/inputs/online-forms.md", "sha256": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c" }, "email-marketing": { "input": "research/inputs/email-marketing.md", "sha256": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "runner": { "input": "research/inputs/claude-runner.md", "sha256": "09715b90efc5c8fbd280f9d9999dfe26744555c78802c3b0b2a4779c15c4e47c" } }, "runs": [ { "id": "codex", "run_id": "codex-2026-09-28", "researcher": "Codex", "model_id": "gpt-6-astra", "model_note": "研究与整合模型：GPT-6 Astra，推理强度 high。", "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:35:14.962471+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "execution": "Primary assistant live web research; manually structured findings. No Claude findings read before this run was saved. Category seeds were chosen by this researcher, so discovery independence is limited.", "hands_on_testing": false, "model_display_name": "GPT-6 Astra", "reasoning_effort": "high", "applies_to": [ "codex-2026-09-28", "integration" ] }, { "id": "claude", "run_id": "claude-2026-09-28", "researcher": "Claude Personal", "model_id": "claude-opus-5-5", "model_note": "主研究模型由 CLI init/result 元数据确认为 claude-opus-5-5；工具辅助模型记录包含 claude-haiku-4-5-20251001。", "auxiliary_model_ids": [ "claude-haiku-4-5-20251001" ], "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:40:57.839849+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "runner": "claude-personal CLI 2.1.281", "runner_input": "research/inputs/claude-runner.md", "runner_input_sha256": "09715b90efc5c8fbd280f9d9999dfe26744555c78802c3b0b2a4779c15c4e47c", "execution": "Separate CLI process with only WebSearch/WebFetch tools; no access to Codex findings. Same category briefs and protocol; discovery methods and tool providers differ.", "duration_ms": 614931, "tool_calls": 56, "hands_on_testing": false } ], "categories": [ { "id": "workflow-automation", "title": "工作流自动化", "question": "Best workflow automation for small teams: no-code operations, visual branching, or developer-controlled workflows?", "candidates": [ "Zapier", "Make", "n8n" ], "audience": "3–10 人的小团队；分别考虑无代码运营、复杂可视化流程、有开发维护能力的团队。", "scenarios": [ "非技术成员维护的常见 SaaS 同步", "多分支与数据转换的可视化自动化", "需要自托管及自定义代码的技术团队" ], "criteria": [ "维护门槛", "目标集成是否可用", "分支与调试能力", "部署控制", "实际计费单位" ], "special_prompt": "prompts/categories/workflow-automation.md", "runs": [ { "id": "codex", "metadata": { "run_id": "codex-2026-09-28", "researcher": "Codex", "model_id": "gpt-6-astra", "model_note": "研究与整合模型：GPT-6 Astra，推理强度 high。", "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:35:14.962471+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "execution": "Primary assistant live web research; manually structured findings. No Claude findings read before this run was saved. Category seeds were chosen by this researcher, so discovery independence is limited.", "hands_on_testing": false, "model_display_name": "GPT-6 Astra", "reasoning_effort": "high", "applies_to": [ "codex-2026-09-28", "integration" ] }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "workflow-automation", "question": "Best workflow automation for small teams: no-code operations, visual branching, or developer-controlled workflows?", "scope_limitations": [ "仅比较三个预先指定候选，未覆盖整个市场。", "研究使用有利益关系的竞争对手材料；未做产品实测。" ], "sources": [ { "id": "S1", "url": "https://blog.n8n.io/free-zapier-alternatives/", "title": "9 Zapier Tools Alternatives", "publisher": "n8n", "publisher_product": "n8n", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Is Zapier the best automation tool?; Make vs Zapier", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S2", "url": "https://zapier.com/blog/n8n-vs-make/", "title": "n8n vs. Make", "publisher": "Zapier", "publisher_product": "Zapier", "relationship": "competing_vendor", "published_at": "2026-03-03", "accessed_at": "2026-09-28", "access_status": "read", "locator": "At a glance; Make is plug-and-play; pricing model", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S3", "url": "https://www.activepieces.com/blog/no-code-automation-unopinionated-extensible", "title": "No-code automation can be unopinionated and extensible", "publisher": "Activepieces", "publisher_product": "Activepieces", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "The Struggle Between Learning Curve and More Features", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" } ], "claims": [ { "id": "C1", "product": "Zapier", "statement": "n8n 和 Activepieces 都将 Zapier 定位为较易上手、面向普通使用者的自动化工具。", "source_ids": [ "S1", "S3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C2", "product": "Zapier", "statement": "n8n 指出 Zapier 的用量计费和套餐功能限制需要特别检查。", "source_ids": [ "S1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "来源中的历史价格不作为当前报价；任务与操作不可直接等价。" }, { "id": "C3", "product": "Make", "statement": "n8n 和 Zapier 都描述 Make 的可视化设计、复杂流程能力及一定学习门槛。", "source_ids": [ "S1", "S2" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C4", "product": "Make", "statement": "Zapier 提醒多步骤、轮询和循环会影响 Make 的用量成本。", "source_ids": [ "S2" ], "kind": "competitor_claim", "status": "unverified", "caveat": "具体计费动作和额度未核实，不据此计算价格优势。" }, { "id": "C5", "product": "n8n", "statement": "Zapier 与 Activepieces 都把 n8n 定位为更面向技术使用者的方案。", "source_ids": [ "S2", "S3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C6", "product": "n8n", "statement": "Zapier 描述 n8n 的自托管选项，同时指出维护与基础设施责任。", "source_ids": [ "S2" ], "kind": "competitor_claim", "status": "unverified", "caveat": "不能把自托管等同于免费、无限或安全性更高；托管版本另算。" } ], "recommendations": [ { "product": "Zapier", "best_for": "非技术成员维护常见 SaaS 同步", "not_for": "仅凭最低标价选择大量多步骤流程", "rationale": "两家对手认可易用定位；先检查实际所需集成与用量。", "claim_ids": [ "C1", "C2" ], "evidence_strength": "moderate", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "Make", "best_for": "愿意学习可视化分支和数据转换的运营团队", "not_for": "不愿承担任何配置学习的使用者", "rationale": "可视化复杂流程是对手承认的特点，费用仍需按真实流程验证。", "claim_ids": [ "C3", "C4" ], "evidence_strength": "moderate", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "n8n", "best_for": "有开发维护能力、需要部署控制的团队", "not_for": "没有运维责任人的自托管项目", "rationale": "技术导向与自托管支持形成试用理由，维护成本不能忽略。", "claim_ids": [ "C5", "C6" ], "evidence_strength": "moderate", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" } ], "conflicts": [ { "topic": "集成数量与计费版本", "source_ids": [ "S1", "S2" ], "description": "旧对比采用 operations 和不同集成数量，新文采用 credits；不把旧数量和现行报价混为一谈。", "resolution": "未解决；不以该项形成确定推荐。" } ], "excluded": [ { "url": "https://blog.n8n.io/free-zapier-alternatives/", "publisher": "n8n", "target": "n8n", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://zapier.com/blog/n8n-vs-make/", "publisher": "Zapier", "target": "Zapier", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://www.activepieces.com/blog/no-code-automation-unopinionated-extensible", "publisher": "Activepieces", "target": "Activepieces", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" } ], "coverage_gaps": [ "未按统一实际流程测量成本、重试及可靠性。", "n8n 自托管的许可范围需要单独核实；不使用来源的 open-source 泛称作为法律结论。" ] } }, { "id": "claude", "metadata": { "run_id": "claude-2026-09-28", "researcher": "Claude Personal", "model_id": "claude-opus-5-5", "model_note": "主研究模型由 CLI init/result 元数据确认为 claude-opus-5-5；工具辅助模型记录包含 claude-haiku-4-5-20251001。", "auxiliary_model_ids": [ "claude-haiku-4-5-20251001" ], "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:40:57.839849+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "runner": "claude-personal CLI 2.1.281", "runner_input": "research/inputs/claude-runner.md", "runner_input_sha256": "09715b90efc5c8fbd280f9d9999dfe26744555c78802c3b0b2a4779c15c4e47c", "execution": "Separate CLI process with only WebSearch/WebFetch tools; no access to Codex findings. Same category briefs and protocol; discovery methods and tool providers differ.", "duration_ms": 614931, "tool_calls": 56, "hands_on_testing": false }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "workflow-automation", "question": "Best workflow automation for small teams: no-code operations, visual branching, or developer-controlled workflows?", "scope_limitations": [ "受众固定为 3–10 人小团队；只比较 Zapier、Make、n8n 三个预设候选，不中途增补。", "证据全部来自竞品厂商的对比页，属于利益相关方的说法；没有做实际测试，也没有逐个核对目标集成的具体 action。", "价格数字都来自竞品转述，币种、计费周期和档位名称各有不一致，不按具体价格排序。", "以后值得加入的候选：Pipedream、Activepieces、Power Automate（本次未研究）。" ], "sources": [ { "id": "W1", "url": "https://zapier.com/blog/n8n-vs-make/", "title": "n8n vs. Make: Which is best? [2026]", "publisher": "Zapier", "publisher_product": "Zapier", "relationship": "n8n 和 Make 的直接竞争对手", "published_at": "2026-03", "accessed_at": "2026-09-28", "access_status": "read", "locator": "易用性、自托管、计费模式、自定义代码各节", "note": "2025-10 首发，2026-03 更新；结论部分在推广 Zapier，已排除。有商业动机。" }, { "id": "W2", "url": "https://zapier.com/blog/n8n-vs-zapier/", "title": "Zapier vs. n8n Comparison: Pricing, Security & Self-Hosting", "publisher": "Zapier", "publisher_product": "Zapier", "relationship": "n8n 的直接竞争对手", "published_at": "2026-09-16", "accessed_at": "2026-09-28", "access_status": "read", "locator": "自托管负担、许可证、学习曲线、FAQ 隐性成本", "note": "开头定位、功能对比表和 FAQ 大多在推广 Zapier，已排除。" }, { "id": "W4", "url": "https://www.make.com/en/compare/make-vs-zapier", "title": "Make vs Zapier in 2026 | Compare features & pricing", "publisher": "Make", "publisher_product": "Make", "relationship": "Zapier 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "线性工作流局限、Canvas、定价、每个应用的 action 深度", "note": "页面只标年份 2026。对 Zapier 易用性和集成广度有让步。" }, { "id": "W5", "url": "https://www.make.com/en/compare/make-vs-n8n", "title": "Make vs N8N in 2026 | Compare features & pricing", "publisher": "Make", "publisher_product": "Make", "relationship": "n8n 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "自托管与技术开销、定价模型、编排、支持", "note": "只标年份 2026；合规认证对比和 Make 自身的价格与功能描述已排除。" }, { "id": "W6", "url": "https://n8n.io/vs/make/", "title": "n8n vs Make", "publisher": "n8n", "publisher_product": "n8n", "relationship": "Make 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Costs、Features Comparison、UI Built for Developers、Self-hosting、Customer Support", "note": "引用了 Make 在 2025-08-27 宣布的计费切换；n8n 自身定位已排除。" }, { "id": "W7", "url": "https://n8n.io/vs/zapier/", "title": "n8n vs Zapier", "publisher": "n8n", "publisher_product": "n8n", "relationship": "Zapier 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "定价、Paths/Webhooks、自定义代码限制、自托管、错误处理", "note": "没有日期；对 Zapier 的低门槛有让步。" } ], "claims": [ { "id": "WC1", "product": "Zapier", "statement": "两个竞品都承认 Zapier 入门门槛低、适合非技术用户做简单自动化（n8n：几分钟就能开始；Make：逐步界面对新手友好）。", "source_ids": [ "W7", "W4" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞品对易用性的称赞不等于可用性测试，而且两家都把它限定在“简单场景”。" }, { "id": "WC2", "product": "Zapier", "statement": "Make 认为 Zapier 的线性模型处理多条件分支、数据迭代和聚合比较吃力；n8n 称 Paths 只在付费档提供，免费档限两步流程，Webhooks 从 Pro 档起。", "source_ids": [ "W4", "W7" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "两家都说复杂分支是 Zapier 的弱项，但档位细节只有 n8n 一家来源，未独立核实。" }, { "id": "WC3", "product": "Zapier", "statement": "Zapier 按 task（每个动作步骤）计费。n8n 称超出套餐的 task 按 1.25 倍计价，单个工作流可能消耗数千个 task；Make 称 Zapier 每个 task 的单价明显高于 Make 每个 credit 的单价。", "source_ids": [ "W7", "W4" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "task、credit、execution 不能直接等价换算；具体价格见冲突项。" }, { "id": "WC4", "product": "Zapier", "statement": "n8n 称 Code by Zapier 有 30 秒执行时间和 256MB 内存上限，且 Zapier 只有云端版本，不能自托管。", "source_ids": [ "W7" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家竞品来源。" }, { "id": "WC5", "product": "Make", "statement": "Zapier 称 Make 是可视化拖拽界面，有学习曲线，但运营、市场、支持等业务人员能上手；Zapier 和 n8n 都说 Make 是全托管云服务、不需要部署。", "source_ids": [ "W1", "W6" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "“全托管”有两家印证；“业务人员能上手”只有 Zapier 一家说。n8n 只承认 Make 适合“简单线性任务”。" }, { "id": "WC6", "product": "Make", "statement": "Make 按每个动作计费（含触发器），过滤、格式化、循环都消耗 credits；AI 或复杂模块按 token、文件大小或处理时间可变消耗，预算难预测。", "source_ids": [ "W1", "W6" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "Make 正从 operations 过渡到 credits（W6 提到 2025-08-27 的公告），当前换算规则未核实。" }, { "id": "WC7", "product": "Make", "statement": "n8n 称 Make 的多分支很难重新合并成一条路径，错误处理要给每个模块单独加 handler 路由，自定义 JS 函数只在 Enterprise 档提供，而且场景始终在云端执行。", "source_ids": [ "W6" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家竞品来源，对“可视化分支与调试”这一评估标准有直接影响。" }, { "id": "WC8", "product": "n8n", "statement": "三家竞品来源都说 n8n 按工作流执行次数计费、与步骤数无关，适合多步骤复杂流程；Make 承认 n8n 自托管社区版没有许可费、不限执行次数，还支持自定义节点、HTTP 请求节点和并行分支。", "source_ids": [ "W1", "W5", "W2" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "“不限执行”只指自托管；云版按执行次数分档，价格未核实。" }, { "id": "WC9", "product": "n8n", "statement": "Zapier 和 Make 都说 n8n 需要较强技术能力（JS 表达式、JSON、API 认证），自托管要自己负责补丁、更新、安全和可用性，还有服务器与人力的隐性成本；Zapier 称它是 fair-code / source-available，不是严格意义的开源；Make 称社区节点存在未审核代码的风险。", "source_ids": [ "W1", "W2", "W5" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "许可证的具体条款没有独立核实；自托管不等于零成本。" } ], "recommendations": [ { "product": "Zapier", "best_for": "场景一：由非技术成员维护的常见 SaaS 之间的简单、线性同步", "not_for": "多分支、迭代聚合的高频流程；需要自托管或自定义代码不受限的团队", "rationale": "两个竞品都承认它门槛低（WC1）；同时两家都指出它按 task 计费、复杂分支受限（WC2、WC3）。", "claim_ids": [ "WC1", "WC2", "WC3", "WC4" ], "evidence_strength": "low", "caveat": "易用性证据来自竞品的让步，不是测试。目标集成里需要的 trigger/action 要逐个到集成页核对，不能看集成总数。实际价格没有核实。" }, { "product": "Make", "best_for": "场景二：由业务侧维护、需要可视化多分支和数据转换、接受云端托管的团队", "not_for": "要求自托管，或需要在非 Enterprise 档写自定义代码的团队；对按步骤计费敏感的高频循环流程", "rationale": "Zapier 称业务人员能上手 Make 的可视化界面（WC5）；但两家竞品都警告按动作/credit 计费会随步骤、循环和 AI 模块增加（WC6），n8n 还指出分支合并和错误处理比较繁琐（WC7）。", "claim_ids": [ "WC5", "WC6", "WC7" ], "evidence_strength": "low", "caveat": "“可视化分支强”的正面证据主要是 Make 自己的说法（已排除），竞品只给了有限的让步；调试体验没有竞品之外的证据。" }, { "product": "n8n", "best_for": "场景三：有开发维护能力、需要自托管、自定义代码或按执行计费的技术团队", "not_for": "没有运维能力的非技术团队；希望完全免维护的团队", "rationale": "Zapier 和 Make 两个竞品家族都承认它的自托管能力、自定义能力和按执行计费（WC8），也都一致指出技术门槛和运维负担（WC9）。这些说法的方向和适用场景吻合。", "claim_ids": [ "WC8", "WC9" ], "evidence_strength": "moderate", "caveat": "竞品之间意见一致，仍不等于独立验证。fair-code 许可对商业使用的限制要自行阅读许可证；自托管的服务器、备份、安全补丁和人力成本要计入总成本。" } ], "conflicts": [ { "topic": "集成数量", "source_ids": [ "W2", "W7", "W5", "W4" ], "description": "Zapier 自称约 9,000+，说 n8n 约 2,000（W2）；n8n 称 Zapier 有 8,000+（W7）；Make 称 Zapier 8,000+、n8n 1,200+（W4、W5）。", "resolution": "未解决。各家口径和日期不同；按协议不按集成数量排序。" }, { "topic": "Zapier 入门付费档价格与名称", "source_ids": [ "W4", "W7" ], "description": "Make 称“Starter $19.99/月，750 tasks”，n8n 称“Pro $19.99/月，750 tasks”；档位名称不一致，也没说明是年付折算还是月付。", "resolution": "未解决；不做价格排序。" }, { "topic": "n8n 云版价格", "source_ids": [ "W5" ], "description": "Make 列出 Starter $20/月（2,500 次执行）、Pro $50/月（10,000 次）、Business $667/月（40,000 次），但没有说明币种（可能是欧元）和计费周期。", "resolution": "只有一家来源，币种和周期未确认，不采用。" }, { "topic": "Make 计费单位", "source_ids": [ "W1", "W6" ], "description": "W6 说 Make 正从 operations 过渡到 credits，W1 已按 credits 描述。", "resolution": "当前换算规则未核实，按“可变消耗”处理。" } ], "excluded": [ { "url": "https://blog.n8n.io/make-vs-zapier/", "publisher": "n8n", "target": "Make、Zapier", "reason": "已读取，但更新日期是 2024-05-22，信息过时，也超出每类 6 个来源的上限。关于 Make 的 Router 和错误处理选项的说法没有作为证据使用。" }, { "url": "https://zapier.com/blog/n8n-vs-make/", "publisher": "Zapier", "target": "Zapier", "reason": "“Which automation platform is right for you?”等结论段在推广 Zapier，属于自我推广。" }, { "url": "https://zapier.com/blog/n8n-vs-zapier/", "publisher": "Zapier", "target": "Zapier", "reason": "Zapier 关于自身 SOC 2 / GDPR 合规和托管优势的表述属于自我描述和合规声明，已排除。" }, { "url": "https://www.make.com/en/compare/make-vs-n8n", "publisher": "Make", "target": "Make、n8n", "reason": "Make 自身的价格（Core $9/月）、3,000+ 集成和 SOC/ISO 认证对比属于自我推广或合规声明，已排除。" }, { "url": "https://n8n.io/vs/zapier/", "publisher": "n8n", "target": "n8n", "reason": "n8n 的自我定位段落（按执行计费、完全代码控制）已排除。" }, { "url": "https://www.datacamp.com/blog/n8n-vs-zapier", "publisher": "DataCamp", "target": "n8n、Zapier", "reason": "中立或第三方评测站，只用于发现线索，不作为排名证据。" }, { "url": "https://parseur.com/blog/zapier-n8n-make", "publisher": "Parseur", "target": "三者", "reason": "与三者的竞争关系不明确，只用于发现线索。" } ], "coverage_gaps": [ "所有价格都没有从现行官方价目核实（按严格政策不能用自述），无法计算小团队的实际月成本。", "没有按具体目标 SaaS 核对 trigger/action 是否可用；集成总数不能说明需要的动作是否存在。", "Make 的“可视化分支和调试”优势缺少竞品方面的正面证据，只有 Zapier 一家的易用性让步。", "重试、轮询频率对 task、credit、execution 消耗的影响，竞品来源都没有覆盖。", "n8n Sustainable Use License 对商业使用的具体限制没有经过竞品之外的核实。" ] } } ], "integration": { "category_id": "workflow-automation", "summary": "三款分别保留为易上手连接、可视化流程和技术团队控制权的试用候选；没有跨场景总冠军。", "entries": [ { "product": "Zapier", "best_for": "非技术人员连接常见 SaaS，优先尽快跑通简单流程。", "not_for": "要求自托管，或复杂分支和高频多步骤任务必须精确控成本。", "rationale": "两轮都提取到对手认可的较低入门门槛；推荐仅限定在简单连接场景。", "caveat": "未验证你的具体连接器、重试机制或当前套餐；对手的价格比较不能代替同一工作负载账单。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C1" }, { "run": "codex", "claim_id": "C2" }, { "run": "claude", "claim_id": "WC1" } ] }, { "product": "Make", "best_for": "愿意学习可视化流程、需要分支和数据转换的运营团队。", "not_for": "希望完全不学习流程概念，或必须自行部署基础设施。", "rationale": "两轮对视觉化流程定位一致。Codex 的正向对手材料较多，Claude 对证据力度更保守。", "caveat": "用量单位和复杂模块消耗未统一；不采纳“只能在 Enterprise 写代码”的未复核限制为当前事实。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C3" }, { "run": "codex", "claim_id": "C4" }, { "run": "claude", "claim_id": "WC5" }, { "run": "claude", "claim_id": "WC6" } ] }, { "product": "n8n", "best_for": "有开发和运维能力，重视自托管选项或定制流程的团队。", "not_for": "无人负责升级、备份、权限和故障恢复的小团队。", "rationale": "两轮都认为技术门槛与控制权相伴；从这个取舍出发列入候选。", "caveat": "未核算托管、人力和安全成本；社区版可用不等于部署免费或适合任何商业用途。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C5" }, { "run": "codex", "claim_id": "C6" }, { "run": "claude", "claim_id": "WC9" } ] } ], "differences": [ "两轮都推荐三个产品，但 Make 的证据强度判断不同；不将推荐一致转成高置信度。", "Claude WC8 写“三家竞品来源”，实际是三个页面、两个厂商（Zapier 与 Make），不是三个独立对手。", "同一 Zapier 对比页被两轮使用，只计一份来源；task、credit、execution 没有统一为相同负载。" ], "next_checks": [ "对同一真实流程统计动作、循环、轮询、失败重试和月执行量。", "核对所需连接器权限与自托管维护能力，完成一次端到端试用。" ] }, "review_notes": [ { "category_id": "workflow-automation", "run": "claude", "claim_id": "WC3", "note": "Make 对自身 credit 的单价描述是自述，整合时排除；task 与 credit 不能直接等价。" }, { "category_id": "workflow-automation", "run": "claude", "claim_id": "WC8", "note": "应读作三个页面、两个发布厂商；不增加为三个独立竞争对手。" } ] }, { "id": "project-management", "title": "项目管理", "question": "Best project management for a 10-person cross-functional team?", "candidates": [ "Asana", "ClickUp", "monday.com" ], "audience": "10 人跨职能团队，比较结构化协作、工具整合、可视化运营三种需求。", "scenarios": [ "清晰的负责人、依赖关系和跨团队协作", "任务、文档与工时整合", "非技术团队的可视化业务流程" ], "criteria": [ "上手与维护成本", "任务和依赖模型", "跨项目可见性", "必需功能的套餐限制", "10 人总成本" ], "special_prompt": "prompts/categories/project-management.md", "runs": [ { "id": "codex", "metadata": { "run_id": "codex-2026-09-28", "researcher": "Codex", "model_id": "gpt-6-astra", "model_note": "研究与整合模型：GPT-6 Astra，推理强度 high。", "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:35:14.962471+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "execution": "Primary assistant live web research; manually structured findings. No Claude findings read before this run was saved. Category seeds were chosen by this researcher, so discovery independence is limited.", "hands_on_testing": false, "model_display_name": "GPT-6 Astra", "reasoning_effort": "high", "applies_to": [ "codex-2026-09-28", "integration" ] }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "project-management", "question": "Best project management for a 10-person cross-functional team?", "scope_limitations": [ "仅比较三个预先指定候选，未覆盖整个市场。", "研究使用有利益关系的竞争对手材料；未做产品实测。" ], "sources": [ { "id": "S1", "url": "https://agiled.app/compare/asana-vs-clickup-vs-monday", "title": "Asana vs ClickUp vs Monday", "publisher": "Agiled", "publisher_product": "Agiled", "relationship": "competing_vendor", "published_at": "2026-04-07", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Quick verdict; task management; pricing; feature table", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S2", "url": "https://help.clickup.com/hc/en-us/articles/40980356364311-ClickUp-vs-Monday-com-Which-project-management-tool-fits-your-team", "title": "ClickUp vs. Monday.com", "publisher": "ClickUp", "publisher_product": "ClickUp", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Introduction; Monday pricing table; pricing verified April 2026", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" } ], "claims": [ { "id": "C1", "product": "Asana", "statement": "Agiled 将 Asana 描述为结构化工作管理和跨项目规划的候选。", "source_ids": [ "S1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C2", "product": "Asana", "statement": "Agiled 同一页对 Goals/Portfolios 所属套餐出现不一致描述。", "source_ids": [ "S1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "开头把它们与 Starter 连用，随后又写需要 Advanced；套餐未定。" }, { "id": "C3", "product": "ClickUp", "statement": "Agiled 将 ClickUp 描述为结合任务、文档、白板与工时的高功能密度方案。", "source_ids": [ "S1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C4", "product": "ClickUp", "statement": "Agiled 认为 ClickUp 配置和学习负担较高。", "source_ids": [ "S1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "不采用其“几周”上手时间及性能投诉作为测量事实。" }, { "id": "C5", "product": "monday.com", "statement": "Agiled 与 ClickUp 都认可 monday.com 的可视化、颜色状态和看板式界面。", "source_ids": [ "S1", "S2" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C6", "product": "monday.com", "statement": "两家对手都提醒功能分层和付费席位门槛。", "source_ids": [ "S1", "S2" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "历史价格未复核；3 席最低并不能证明 10 人的完整费用。" } ], "recommendations": [ { "product": "Asana", "best_for": "重视清晰职责和结构化协作的跨职能团队", "not_for": "必须按最低套餐获得所有高级规划功能的团队", "rationale": "结构化定位有对手支持，但只有一个出版方，且存在套餐矛盾。", "claim_ids": [ "C1", "C2" ], "evidence_strength": "low", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "ClickUp", "best_for": "愿意指定管理员整合任务、文档和工时的团队", "not_for": "希望几乎零配置直接使用的团队", "rationale": "功能整合可作为试用假设，配置成本需要实际验证。", "claim_ids": [ "C3", "C4" ], "evidence_strength": "low", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "monday.com", "best_for": "偏好表格与可视化业务流程的非技术团队", "not_for": "尚未确认必需功能所属套餐就采购的团队", "rationale": "界面定位得到两家对手认可；10 人应另做套餐核价。", "claim_ids": [ "C5", "C6" ], "evidence_strength": "moderate", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" } ], "conflicts": [ { "topic": "Asana 套餐自相矛盾", "source_ids": [ "S1" ], "description": "Agiled 对 Goals/Portfolios 的 Starter 与 Advanced 归属不一致；拒绝沿用其起价推导团队成本。", "resolution": "未解决；不以该项形成确定推荐。" } ], "excluded": [ { "url": "https://agiled.app/compare/asana-vs-clickup-vs-monday", "publisher": "Agiled", "target": "Agiled", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://help.clickup.com/hc/en-us/articles/40980356364311-ClickUp-vs-Monday-com-Which-project-management-tool-fits-your-team", "publisher": "ClickUp", "target": "ClickUp", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" } ], "coverage_gaps": [ "Asana 与 ClickUp 的有效推荐证据各仅一个竞争出版方。", "未确认 10 人席位、AI 附加费与高级功能后的等价总成本。" ] } }, { "id": "claude", "metadata": { "run_id": "claude-2026-09-28", "researcher": "Claude Personal", "model_id": "claude-opus-5-5", "model_note": "主研究模型由 CLI init/result 元数据确认为 claude-opus-5-5；工具辅助模型记录包含 claude-haiku-4-5-20251001。", "auxiliary_model_ids": [ "claude-haiku-4-5-20251001" ], "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:40:57.839849+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "runner": "claude-personal CLI 2.1.281", "runner_input": "research/inputs/claude-runner.md", "runner_input_sha256": "09715b90efc5c8fbd280f9d9999dfe26744555c78802c3b0b2a4779c15c4e47c", "execution": "Separate CLI process with only WebSearch/WebFetch tools; no access to Codex findings. Same category briefs and protocol; discovery methods and tool providers differ.", "duration_ms": 614931, "tool_calls": 56, "hands_on_testing": false }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "project-management", "question": "Best project management for a 10-person cross-functional team?", "scope_limitations": [ "受众是 10 人跨职能团队；只比较 Asana、ClickUp、monday.com 的项目管理产品，不包括 monday CRM/dev 等相邻产品。", "证据全部来自竞品厂商；上手时间和性能说法都没有实测。", "10 人总成本没法可靠计算：各来源的计费周期、席位包规则和 AI 加购说法不一致。", "以后值得加入：Wrike、Notion、Linear（跨职能场景下，本次未研究）。" ], "sources": [ { "id": "P1", "url": "https://clickup.com/blog/monday-vs-asana-vs-clickup-small-business-pricing/", "title": "Monday vs. Asana vs. ClickUp Pricing: Which Is Best Value?", "publisher": "ClickUp", "publisher_product": "ClickUp", "relationship": "Asana 和 monday.com 的直接竞争对手", "published_at": "2026-02-21", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Monday.com 定价结构、Asana 定价结构、Plan Tiers & Features", "note": "ClickUp 自身免费档、Brain、Super Agents 等段落属于自我推广，已排除。" }, { "id": "P2", "url": "https://monday.com/blog/project-management/clickup-vs-asana-how-they-compare-and-which-to-choose/", "title": "ClickUp vs Asana (2026): Features, Pricing, And AI Compared", "publisher": "monday.com", "publisher_product": "monday work management", "relationship": "ClickUp 和 Asana 的直接竞争对手", "published_at": "2026-06-24", "accessed_at": "2026-09-28", "access_status": "read", "locator": "定价对比表、各产品优劣段", "note": "文末定位 monday.com 为“第三选择”，已排除。" }, { "id": "P3", "url": "https://asana.com/compare/asana-vs-clickup", "title": "Asana vs. ClickUp", "publisher": "Asana", "publisher_product": "Asana", "relationship": "ClickUp 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "设置复杂度、AI、自动化限制、移动端各段", "note": "页脚提到 2025 年 12 月，没有正式发布日期。" }, { "id": "P4", "url": "https://asana.com/compare/asana-vs-monday-com", "title": "Asana vs. Monday.com", "publisher": "Asana", "publisher_product": "Asana", "relationship": "monday.com 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "可视化、套餐结构、席位块、跨项目可见性", "note": "案例研究、G2/Gartner 奖项段属于自我推广，已排除。" }, { "id": "P5", "url": "https://clickup.com/blog/asana-vs-monday/", "title": "Asana Vs Monday: Project Management Tool Comparison", "publisher": "ClickUp", "publisher_product": "ClickUp", "relationship": "Asana 和 monday.com 的直接竞争对手（与 P1 同属 ClickUp 家族）", "published_at": "2024-04-07", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Asana 优缺点、Monday 优缺点、价格表", "note": "内容已过时（2024）；与 P1 算作同一证据家族，只用于可视化让步和冲突比较。" }, { "id": "P6", "url": "https://monday.com/blog/project-management/clickup-vs-asana-vs-monday-work-management/", "title": "ClickUp vs Asana vs Monday Work Management Comparison 2027", "publisher": "monday.com", "publisher_product": "monday work management", "relationship": "ClickUp 和 Asana 的直接竞争对手（与 P2 同属 monday 家族）", "published_at": "2026-01-30", "accessed_at": "2026-09-28", "access_status": "read", "locator": "ClickUp 评估、Asana 评估、跨项目可见性", "note": "标题写“2027”，晚于研究截止日期；以正文日期 2026-01-30 为准。作者 Ben Kazinik。" } ], "claims": [ { "id": "PC1", "product": "Asana", "statement": "monday.com 称 Asana 界面清晰、结构化，非技术团队能较快上手，任务负责人和优先级清楚。", "source_ids": [ "P2", "P6" ], "kind": "competitor_claim", "status": "unverified", "caveat": "P2 和 P6 都是 monday.com 发布，只算一个证据家族；上手速度是竞品的判断，不是测量结果。" }, { "id": "PC2", "product": "Asana", "statement": "ClickUp 和 monday.com 都说 Asana 的付费档已经包含 AI，不用另外加购。", "source_ids": [ "P1", "P2" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "AI 在各档的具体额度没有核实。" }, { "id": "PC3", "product": "Asana", "statement": "付费档最少 2 个席位；Portfolios、工作负载管理和原生工时追踪要到较高档（Advanced 或以上）才有；Starter 起价被引为 $10.99/用户/月。", "source_ids": [ "P1", "P2", "P6" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "$10.99 没有说明是年付还是月付；跨项目可见性（Portfolios）取决于档位，对“跨项目可见性”标准影响很大。" }, { "id": "PC4", "product": "Asana", "statement": "monday.com 称 Asana 层级结构刚性、难以建模复杂业务流程，高级报表和资源管理依赖更高档或 Tableau/PowerBI 等第三方工具。", "source_ids": [ "P6" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家竞品来源。" }, { "id": "PC5", "product": "ClickUp", "statement": "Asana 承认 ClickUp 功能覆盖面广（思维导图、不限存储、原生工时追踪）；monday.com 称 ClickUp 把项目管理、文档、聊天、白板整合在一起，免费档不限成员。", "source_ids": [ "P3", "P2" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "功能多不代表容易推行；免费档的存储等限制要另行核对。" }, { "id": "PC6", "product": "ClickUp", "statement": "Asana 和 monday.com 都说 ClickUp 设置量大、学习曲线陡、新用户容易被压倒；monday.com 还称数据量大时会有卡顿。", "source_ids": [ "P3", "P2", "P6" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "性能说法没有测量数据，不当作事实。" }, { "id": "PC7", "product": "ClickUp", "statement": "monday.com 称 ClickUp 的 AI（Brain MAX）要另外加购，约 $9/用户/月；Unlimited 档起价 $7/用户/月；Asana 称 ClickUp 的自动化和集成限制在 Unlimited 档。", "source_ids": [ "P2", "P3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "加购价格只有一家来源，计费周期未说明。" }, { "id": "PC8", "product": "monday.com", "statement": "Asana 承认 monday.com 强调可视化仪表盘、定制和可视化自动化规则；ClickUp（2024）称它的视图选项最多。", "source_ids": [ "P4", "P5" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "P5 已过时；可视化是否便于非技术团队维护属于推断。" }, { "id": "PC9", "product": "monday.com", "statement": "ClickUp 称 monday.com 按席位包计费（最少 3 席，5/10/15 档），时间线/日历要 Standard 档及以上，AI 助手和原生工时追踪在 Pro 档；Asana 也批评它的“刚性席位块”和功能按档限制。", "source_ids": [ "P1", "P4" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "10 人正好对应一个席位包，但各档价格和年付要求没有核实。" } ], "recommendations": [ { "product": "Asana", "best_for": "场景一：需要清晰的负责人、依赖关系和跨团队协作，并希望 AI 已含在订阅内的团队", "not_for": "需要在低档就用上 Portfolios 或原生工时的团队；需要高度自定义业务流程建模的团队", "rationale": "两个竞品都承认 AI 已包含（PC2），monday.com 称它结构清晰（PC1）；但跨项目可见性要到较高档（PC3）。", "claim_ids": [ "PC1", "PC2", "PC3", "PC4" ], "evidence_strength": "low", "caveat": "“易上手”只有 monday.com 一个家族的说法；依赖关系在哪一档可用没有从竞品处得到明确的当前证据。" }, { "product": "ClickUp", "best_for": "场景二：想把任务、文档、工时放进一个工具，并愿意投入初期配置的团队", "not_for": "希望零配置、马上上手的非技术团队", "rationale": "Asana 和 monday.com 都承认它整合度高、有原生工时（PC5），也都警告设置复杂（PC6）。", "claim_ids": [ "PC5", "PC6", "PC7" ], "evidence_strength": "moderate", "caveat": "AI 加购会增加 10 人成本；竞品意见一致，仍不等于独立验证。" }, { "product": "monday.com", "best_for": "场景三：非技术团队用可视化看板运营业务流程", "not_for": "预算紧、需要在低档用上时间线、工时和 AI 的团队", "rationale": "Asana 和 ClickUp 都承认它可视化灵活（PC8），同时都批评席位包和功能按档限制（PC9）。", "claim_ids": [ "PC8", "PC9" ], "evidence_strength": "low", "caveat": "依赖关系的当前支持情况有冲突（见 conflicts）；10 人总价未核实。" } ], "conflicts": [ { "topic": "Asana 免费档用户上限", "source_ids": [ "P5", "P1", "P2" ], "description": "P5（2024）说最多 15 人；P1 和 P2（2026）说 2 人或 1–2 人。", "resolution": "按日期采用较新的 2026 说法，P5 视为过时；当前值仍未独立核实。" }, { "topic": "monday.com 依赖关系", "source_ids": [ "P5", "P6" ], "description": "P5（2024）说 monday 没有原生依赖、要用列变通；P6（monday 自己发布）描述了跨板依赖能力，属于自述，已排除。", "resolution": "未解决；P5 已过时，缺少当前的竞品证据。" }, { "topic": "价格及计费周期", "source_ids": [ "P5", "P2" ], "description": "P5 列出 2024 年年付价（Asana $10.99/$24.99，monday $8/$16）；P2 给出 Asana $10.99、ClickUp $7，但没有说明周期。", "resolution": "不计算 10 人总价。" }, { "topic": "集成数量", "source_ids": [ "P2", "P4", "P5" ], "description": "Asana 集成数被说成 270+（P2）或 150+（P5，2024）；monday 被说成 200+（P4）或约 50（P5）。", "resolution": "未解决；不按数量排序。" } ], "excluded": [ { "url": "https://clickup.com/blog/monday-vs-asana-vs-clickup-small-business-pricing/", "publisher": "ClickUp", "target": "ClickUp", "reason": "ClickUp 免费档、Brain、Super Agents、“无席位下限”等自我推广段已排除。" }, { "url": "https://asana.com/compare/asana-vs-monday-com", "publisher": "Asana", "target": "Asana、monday.com", "reason": "Asana 的 ROI 案例和奖项属于自我推广；表格中“monday.com 不存在某功能”这类说法属于从沉默推断缺失，不采用。" }, { "url": "https://monday.com/blog/project-management/clickup-vs-asana-vs-monday-work-management/", "publisher": "monday.com", "target": "monday.com", "reason": "Work OS、AI Blocks、Portfolio Risk Insights 等自我描述已排除；标题“2027”晚于截止日期。" }, { "url": "https://medium.com/@theo-james/i-tested-all-3-clickup-vs-monday-vs-asana-26869c26cd4b", "publisher": "Medium 个人作者", "target": "三者", "reason": "中立或个人评测，只用于发现线索。" }, { "url": "https://zapier.com/blog/clickup-vs-asana/", "publisher": "Zapier", "target": "ClickUp、Asana", "reason": "Zapier 不是项目管理竞品，关系属于相邻或集成方，本次没有读取。" } ], "coverage_gaps": [ "没有核实 10 人的实际总成本：席位包、年付承诺、AI 加购都不清楚。", "依赖关系在各产品哪一档可用，缺少当前竞品证据（monday 的说法来自 2024 年的过时来源）。", "Asana 的“易上手”只有 monday.com 一个家族的证据。", "三者的实际学习时间和性能都没有测量证据。", "ClickUp 对 Asana 和 monday 的有效当前来源只有 P1 一篇（P5 已过时）。" ] } } ], "integration": { "category_id": "project-management", "summary": "按团队协作习惯拆分候选：结构化任务、功能集中、可视化运营。十人团队的准确总价仍未核实。", "entries": [ { "product": "Asana", "best_for": "重视负责人、任务结构与跨项目协作的十人团队。", "not_for": "需要高度自由的数据模型，或要求低档位保证全部资源规划能力。", "rationale": "两轮方向一致。Codex 的直接依据主要来自 Agiled；Claude 增补 monday.com 描述，但整合阶段重读该页超时，未提升证据强度。", "caveat": "Agiled 对 Goals/Portfolios 套餐有内部不一致；Claude 关于 AI 包含范围的说法未作为推荐依据。", "agreement_label": "两轮方向一致 · 新来源待复读", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C1" }, { "run": "codex", "claim_id": "C2" } ] }, { "product": "ClickUp", "best_for": "愿意投入配置时间，希望集中任务、文档与工时功能的团队。", "not_for": "首要要求是最少配置和最短培训时间。", "rationale": "Agiled 的功能密度描述，与 Asana 对 ClickUp 功能广度、配置负担的描述相合。", "caveat": "没有实测性能或迁移成本；不把对手的卡顿指控视为结论，也不采纳未经核对的 AI 加购价。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C3" }, { "run": "codex", "claim_id": "C4" }, { "run": "claude", "claim_id": "PC5" }, { "run": "claude", "claim_id": "PC6" } ] }, { "product": "monday.com", "best_for": "围绕状态看板、可视化仪表盘和定制列管理运营。", "not_for": "无法接受功能分档和潜在席位包成本，或尚不清楚跨项目要求。", "rationale": "Codex 两家来源与 Claude 的 Asana 来源都支持可视化定位。", "caveat": "2024 年比较与当前资料可能冲突；当前功能是否在你的套餐内仍未知。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C5" }, { "run": "codex", "claim_id": "C6" }, { "run": "claude", "claim_id": "PC8" } ] } ], "differences": [ "Claude 找到更多项目管理厂商互评；仅重读通过的相关段落进入整合依据。", "ClickUp 2024 年文章对 monday.com 功能缺失的描述不作为当前否定证据。", "同一厂商的多篇文章不是多个独立观点；Asana 的具体 AI 范围和三个产品总价均未解决。" ], "next_checks": [ "用同一十人配置核对月付、年付、席位包、AI、自动化和报表档位。", "用一个真实跨部门项目试用；核查依赖、权限、访客、导出和迁移。" ] }, "review_notes": [] }, { "id": "meeting-notes", "title": "AI 会议纪要", "question": "Best AI meeting notes for solo users and small client-facing teams?", "candidates": [ "Otter", "Fireflies", "Fathom" ], "audience": "个人知识工作者和 5 人客户沟通团队；分别考虑个人记录、实时协作和 CRM 跟进。", "scenarios": [ "个人会议记录与回看", "实时转录及协作", "销售或客户成功的会后跟进" ], "criteria": [ "会议平台和语言适配", "捕获方式", "转录与摘要的区别", "团队共享和 CRM", "用量限制与数据处理未知项" ], "special_prompt": "prompts/categories/meeting-notes.md", "runs": [ { "id": "codex", "metadata": { "run_id": "codex-2026-09-28", "researcher": "Codex", "model_id": "gpt-6-astra", "model_note": "研究与整合模型：GPT-6 Astra，推理强度 high。", "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:35:14.962471+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "execution": "Primary assistant live web research; manually structured findings. No Claude findings read before this run was saved. Category seeds were chosen by this researcher, so discovery independence is limited.", "hands_on_testing": false, "model_display_name": "GPT-6 Astra", "reasoning_effort": "high", "applies_to": [ "codex-2026-09-28", "integration" ] }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "meeting-notes", "question": "Best AI meeting notes for solo users and small client-facing teams?", "scope_limitations": [ "仅比较三个预先指定候选，未覆盖整个市场。", "研究使用有利益关系的竞争对手材料；未做产品实测。" ], "sources": [ { "id": "S1", "url": "https://fireflies.ai/blog/fireflies-vs-otter", "title": "Fireflies vs. Otter", "publisher": "Fireflies", "publisher_product": "Fireflies", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Choose Otter; Has Otter Caught Up?", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S2", "url": "https://fireflies.ai/blog/fireflies-vs-fathom", "title": "Fireflies vs. Fathom", "publisher": "Fireflies", "publisher_product": "Fireflies", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Side-by-side; Pricing", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S3", "url": "https://otter.ai/blog/otter-vs-fireflies", "title": "Otter vs. Fireflies", "publisher": "Otter", "publisher_product": "Otter", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Our verdict; free tier FAQ", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S4", "url": "https://www.fathom.ai/vs/fireflies", "title": "Fathom vs. Fireflies", "publisher": "Fathom", "publisher_product": "Fathom", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Workflow maintenance; CRM outputs; bot-free FAQs", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S5", "url": "https://tldv.io/blog/honest-review-of-fathom/", "title": "Honest Review of Fathom AI", "publisher": "tl;dv", "publisher_product": "tl;dv", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "TL;DR; first impressions", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" } ], "claims": [ { "id": "C1", "product": "Otter", "statement": "Fireflies 承认 Otter 的个人转录定位及免费使用空间。", "source_ids": [ "S1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C2", "product": "Otter", "statement": "Fireflies 认为 Otter 的语言覆盖和跨系统整合广度较窄。", "source_ids": [ "S1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "未验证语言质量；不复述准确率与隐私指控。" }, { "id": "C3", "product": "Fireflies", "statement": "Otter 承认 Fireflies 具有较广语言覆盖与集成目录。", "source_ids": [ "S3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C4", "product": "Fireflies", "statement": "Fathom 指出 Fireflies 的工作流配置和维护负担；Otter 提醒深度 CRM 等能力存在套餐限制。", "source_ids": [ "S3", "S4" ], "kind": "competitor_claim", "status": "unverified", "caveat": "这是两种不同限制，不能当成同一个缺点的交叉验证。" }, { "id": "C5", "product": "Fathom", "statement": "Fireflies 描述 Fathom 的个人免费记录空间；tl;dv 认可其安静记录和整理会议的基本用途。", "source_ids": [ "S2", "S5" ], "kind": "competitor_claim", "status": "unverified", "caveat": "两者支持用途，不证明所有 AI 摘要及团队功能都无限。" }, { "id": "C6", "product": "Fathom", "statement": "tl;dv 对摘要灵活性评价保留，Fireflies 的比较也区分个人免费能力与团队付费能力。", "source_ids": [ "S2", "S5" ], "kind": "competitor_claim", "status": "unverified", "caveat": "不能把对手的体验判断当成摘要质量基准测试。" } ], "recommendations": [ { "product": "Otter", "best_for": "希望先试用个人转录的使用者", "not_for": "未经语言与平台验证就直接采用的多语言团队", "rationale": "仅作个人转录试用候选；实时协作需求在本轮缺少充分对手证据。", "claim_ids": [ "C1", "C2" ], "evidence_strength": "low", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "Fireflies", "best_for": "需要检查 CRM 路由及较广集成覆盖的客户沟通团队", "not_for": "没有人维护映射、权限和工作流的团队", "rationale": "对手承认广覆盖；具体 CRM 字段和套餐须试验。", "claim_ids": [ "C3", "C4" ], "evidence_strength": "low", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "Fathom", "best_for": "优先考虑个人会议记录与回看的使用者", "not_for": "把个人免费计划当成完整团队方案的采购", "rationale": "两家对手描述了基本记录用途；摘要质量及团队边界未测。", "claim_ids": [ "C5", "C6" ], "evidence_strength": "low", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" } ], "conflicts": [ { "topic": "Fireflies 免费存储描述", "source_ids": [ "S3", "S4" ], "description": "对手材料与比较表的存储、捕获模式描述可能对应不同时间/方案；本轮不发布定量容量结论。", "resolution": "未解决；不以该项形成确定推荐。" }, { "topic": "Fathom 页面捕获描述内部不一致", "source_ids": [ "S4" ], "description": "页面一处称 Fireflies 没有同类无机器人捕获，后面的 FAQ 又以“like Fireflies”提问，不能据此判定功能缺失。", "resolution": "未解决；不以该项形成确定推荐。" } ], "excluded": [ { "url": "https://fireflies.ai/blog/fireflies-vs-otter", "publisher": "Fireflies", "target": "Fireflies", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://fireflies.ai/blog/fireflies-vs-fathom", "publisher": "Fireflies", "target": "Fireflies", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://otter.ai/blog/otter-vs-fireflies", "publisher": "Otter", "target": "Otter", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://www.fathom.ai/vs/fireflies", "publisher": "Fathom", "target": "Fathom", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://tldv.io/blog/honest-review-of-fathom/", "publisher": "tl;dv", "target": "tl;dv", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" } ], "coverage_gaps": [ "Otter 正面定位主要来自 Fireflies，缺少第二个有效对手来源。", "没有实测中文、混合语言或会议平台质量。", "录音同意、保留策略和数据训练政策不能用对手指控作结论。" ] } }, { "id": "claude", "metadata": { "run_id": "claude-2026-09-28", "researcher": "Claude Personal", "model_id": "claude-opus-5-5", "model_note": "主研究模型由 CLI init/result 元数据确认为 claude-opus-5-5；工具辅助模型记录包含 claude-haiku-4-5-20251001。", "auxiliary_model_ids": [ "claude-haiku-4-5-20251001" ], "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:40:57.839849+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "runner": "claude-personal CLI 2.1.281", "runner_input": "research/inputs/claude-runner.md", "runner_input_sha256": "09715b90efc5c8fbd280f9d9999dfe26744555c78802c3b0b2a4779c15c4e47c", "execution": "Separate CLI process with only WebSearch/WebFetch tools; no access to Codex findings. Same category briefs and protocol; discovery methods and tool providers differ.", "duration_ms": 614931, "tool_calls": 56, "hands_on_testing": false }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "meeting-notes", "question": "Best AI meeting notes for solo users and small client-facing teams?", "scope_limitations": [ "受众是个人知识工作者和 5 人客户沟通团队；只比较 Otter、Fireflies、Fathom。", "证据全部来自竞品（包括第四方竞品 Granola）；没有测试转录质量，也没有测普通话或中英混合效果。", "转录准确率、同意录音的合法性、训练数据使用、合规、隐私优劣都列为未解决的前提，不作为事实。", "没有找到 Otter 发布的、关于 Fireflies 或 Fathom 的可读对比页。", "以后值得加入：tl;dv、Granola、飞书妙记/腾讯会议纪要（中文场景）。" ], "sources": [ { "id": "M1", "url": "https://www.fathom.ai/vs/fireflies", "title": "Fathom vs. Fireflies: AI Summaries & Faster ROI (2026)", "publisher": "Fathom", "publisher_product": "Fathom", "relationship": "Fireflies 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Fireflies 弱点、bot-free 录制说明", "note": "只标 2026；G2 评分和 ROI 月数已排除。" }, { "id": "M2", "url": "https://fireflies.ai/blog/fireflies-vs-otter", "title": "Fireflies vs. Otter (2026): Honest AI Assistant Comparison", "publisher": "Fireflies.ai", "publisher_product": "Fireflies", "relationship": "Otter 的直接竞争对手", "published_at": "2026-06-29", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Otter 自 2023 年以来的改进、平台与集成、定价（月付）", "note": "准确率百分比、训练数据和诉讼说法已排除。" }, { "id": "M3", "url": "https://www.granola.ai/blog/meeting-note-tool-pricing-granola-vs-fireflies-fathom-otter", "title": "Meeting note tool pricing: Granola vs. Fireflies vs. Fathom vs. Otter", "publisher": "Granola", "publisher_product": "Granola", "relationship": "三者共同的竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Fireflies、Fathom、Otter 各自的定价、免费档、捕获方式段", "note": "页面只写“March 20”，年份不详；推广 Granola 的结论段已排除。" }, { "id": "M4", "url": "https://fireflies.ai/blog/fireflies-vs-fathom", "title": "Fireflies vs. Fathom", "publisher": "Fireflies.ai", "publisher_product": "Fireflies", "relationship": "Fathom 的直接竞争对手", "published_at": "2026-07-11", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Fathom 优点、平台与捕获限制、CRM 集成、年付价格", "note": "“Verdict at a Glance”和“Fathom 不参与竞争的产品”等段属于自我推广，已排除。" }, { "id": "M5", "url": "https://www.fathom.ai/vs/otter", "title": "Otter vs. Fathom: AI Summaries & Support Compared (2026)", "publisher": "Fathom", "publisher_product": "Fathom", "relationship": "Otter 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Full Video Recordings 节、CRM 同步、Search", "note": "对 Otter 没有任何明确让步；G2 指标已排除。" } ], "claims": [ { "id": "MC1", "product": "Otter", "statement": "Fireflies 和 Granola 都说 Otter 免费档每月 300 分钟转录；Granola 另称单次会话上限 30 分钟。Fireflies 承认这个免费额度“更慷慨”。", "source_ids": [ "M2", "M3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "30 分钟单次上限只有一家来源。" }, { "id": "MC2", "product": "Otter", "statement": "Fireflies 承认 Otter 近年扩展到 34+ 个集成（包括 Salesforce、HubSpot、Zapier、Gmail），推出了 AI Agents，支持 Zoom、Google Meet、Teams、Webex 等；Granola 称 OtterPilot 会按日历自动入会。", "source_ids": [ "M2", "M3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "集成深度（例如是否写入 CRM 字段）没有核实；Fathom 称 Otter 推送到 CRM 的只是通用摘要（M5）。" }, { "id": "MC3", "product": "Otter", "statement": "两家竞品都说 Otter 支持的语言有限（Fireflies：6 种语言变体；Granola：约 4 种）。Fathom 称 Otter 原生只录音频、不录视频。", "source_ids": [ "M2", "M3", "M5" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "语言数量有冲突；有没有中文支持不详。只录音频这一条只有一家来源。" }, { "id": "MC4", "product": "Fireflies", "statement": "Granola 称 Fireflies 会用可见的 bot 入会、也有可选的无 bot 模式，Business 档起支持 Salesforce/HubSpot 同步，并提供面向销售的会话分析；Fathom 也承认 Fireflies 有无 bot 录制。", "source_ids": [ "M3", "M1" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "两家都印证了无 bot 模式；CRM 同步在哪一档只有 Granola 一家来源。“100+ 语言”是语言数量，不代表中文质量。" }, { "id": "MC5", "product": "Fireflies", "statement": "Granola 称 Fireflies 免费档是累计 800 分钟存储（不是按月），另有少量转录和 AI credits，存储很快会用满；功能分档限制较多。", "source_ids": [ "M3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家来源，发布年份不详。" }, { "id": "MC6", "product": "Fireflies", "statement": "Fathom 称 Fireflies 输出偏向原始转录和冗余内容，需要配置工作流、调优和管理维护。", "source_ids": [ "M1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家来源，而且是主观判断。" }, { "id": "MC7", "product": "Fathom", "statement": "Fireflies 和 Granola 都说 Fathom 免费档不限录制次数；Fireflies 称转录和存储也不限，Granola 称每月只有 5 次 AI 摘要。Fireflies 称 Fathom 设置快、摘要质量扎实。", "source_ids": [ "M4", "M3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "“不限录制”有两家印证；免费档 AI 摘要的额度说法不一，属于冲突。" }, { "id": "MC8", "product": "Fathom", "statement": "Fireflies 称 Fathom 只支持 Zoom、Google Meet、Teams，没有移动端，不能转录上传的音视频文件，也没有实时建议；Granola 称 Salesforce/HubSpot 同步要 Team 档才有。", "source_ids": [ "M4", "M3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "每个细节都只有一家来源；不能从一方没提就推断某功能缺失。" }, { "id": "MC9", "product": "Fathom", "statement": "Granola 称 Fathom 以名为“Fathom Notetaker”的可见 bot 入会。", "source_ids": [ "M3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "Fathom 自己说有 bot-free 模式（自述，已排除），见冲突项。" } ], "recommendations": [ { "product": "Fathom", "best_for": "场景一：主要用 Zoom、Meet 或 Teams 的个人，想要低成本记录并回看会议", "not_for": "需要移动端、上传文件转录、免费就要 CRM 同步的用户；中文或混合语言会议（没有证据）", "rationale": "两家竞品都说它免费档不限录制（MC7）；局限在平台范围和 CRM 档位（MC8）。", "claim_ids": [ "MC7", "MC8", "MC9" ], "evidence_strength": "low", "caveat": "免费档 AI 摘要额度有冲突；入会 bot 需要征得参会方同意，这属于未解决的前提。" }, { "product": "Fireflies", "best_for": "场景三：5 人销售或客户成功团队，需要把会议内容同步到 CRM，并希望可选无 bot 捕获", "not_for": "只想用免费档长期存档的个人；不愿做配置维护的团队", "rationale": "两家竞品都承认它有无 bot 模式（MC4），Granola 提到它有 CRM 同步和销售分析；限制是免费档存储是累计上限（MC5），以及 Fathom 所说的维护负担（MC6）。", "claim_ids": [ "MC4", "MC5", "MC6" ], "evidence_strength": "low", "caveat": "CRM 档位和付费价格没有核实；具体要写入哪些 CRM 字段需要试用确认。" } ], "conflicts": [ { "topic": "Fathom 付费档名称与价格", "source_ids": [ "M3", "M4" ], "description": "Granola：Premium 约 $15（年付）/ $19（月付），Team Edition 约 $19 / $29，Team Pro 约 $29 / $39。Fireflies（年付）：Premium $16，Team $15（最少 2 席），Business $25。", "resolution": "未解决；可能是套餐改版造成的，不做价格比较。" }, { "topic": "Otter 语言支持数量", "source_ids": [ "M2", "M3" ], "description": "Fireflies 说 6 种语言变体，Granola 说约 4 种。", "resolution": "未解决；两家都说“有限”，对中文支持都没有证据。" }, { "topic": "Otter 付费价格", "source_ids": [ "M2", "M3" ], "description": "Pro 月付 $16.99 两家一致；Granola 另给出年付约 $8.33、Business $20/$30，Fireflies 给出 Business 月付 $30。", "resolution": "月付数字基本一致，但没有核实，不排序。" }, { "topic": "Fathom 捕获方式", "source_ids": [ "M3", "M1" ], "description": "Granola 说 Fathom 用可见 bot；Fathom 自己的页面说有 bot-free 模式（Mac），属于自述、不采用。", "resolution": "未解决；需要用户自行确认。" }, { "topic": "Fathom 免费档 AI 摘要", "source_ids": [ "M3", "M4" ], "description": "Granola 说每月 5 次 AI 摘要；Fireflies 只说录制、转录、存储不限，没提摘要上限。", "resolution": "不从沉默推断；按“可能有摘要上限”处理。" } ], "excluded": [ { "url": "https://fireflies.ai/blog/fireflies-vs-otter", "publisher": "Fireflies.ai", "target": "Otter", "reason": "“Otter 默认用数据训练 AI”和联邦集体诉讼、准确率 85% 对比 99%、HIPAA 等属于隐私、法律、准确率和合规断言，按协议不当作事实。" }, { "url": "https://fireflies.ai/blog/fireflies-vs-fathom", "publisher": "Fireflies.ai", "target": "Fireflies", "reason": "Fireflies 自称 99%/95% 准确率、100+ 集成等属于自我推广。" }, { "url": "https://www.fathom.ai/vs/fireflies", "publisher": "Fathom", "target": "Fathom、Fireflies", "reason": "G2 评分、ROI 月数（4 个月对比 11 个月）和 Fathom 自身价格、捕获方式都属于自述或营销指标。" }, { "url": "https://www.fathom.ai/vs/otter", "publisher": "Fathom", "target": "Fathom", "reason": "“Switch to Fathom”号召和用户证言属于自我推广。" }, { "url": "https://www.granola.ai/blog/meeting-note-tool-pricing-granola-vs-fireflies-fathom-otter", "publisher": "Granola", "target": "Granola", "reason": "结论段推荐 Granola，属于自我推广。" }, { "url": "https://www.index.dev/blog/otter-vs-fireflies-vs-fathom-ai-meeting-notes-comparison", "publisher": "index.dev", "target": "三者", "reason": "中立或第三方评测，只用于发现线索（包括“Otter 实时协作最强”这类说法）。" } ], "coverage_gaps": [ "场景二（实时转录和协作）不给推荐：没有读到任何竞品页面对 Otter 或其他候选的实时协作能力提供证据，这方面的说法都来自中立站的搜索摘要。", "Otter 没有一个有效的正面推荐场景：竞品只承认它的免费分钟数和集成扩展，不足以支撑推荐。", "普通话和中英混合会议的质量没有任何相关证据，不给多语言推荐。", "数据处理、训练用途、录音同意、合规等都是未解决的前提，使用前要自行审查各产品的条款。", "付费档价格有冲突，也没有核实；上传文件转录的额度只覆盖了 Fathom 的反面说法。", "没有找到 Otter 发布的对比页，所以 Fireflies 和 Fathom 缺少来自 Otter 的视角。" ] } } ], "integration": { "category_id": "meeting-notes", "summary": "Fathom 与 Fireflies 的场景方向获得两轮支持；Otter 仅保留为补充考察项。暂不回答谁的中文识别或实时协作最好。", "entries": [ { "product": "Fathom", "best_for": "个人以会后回顾和会议记录为主，先评估基础记录用途。", "not_for": "未经试用就要求移动端、文件上传、无机器人录制或全部 AI 摘要免费。", "rationale": "两轮都找到对手对个人记录用途与免费记录空间的认可。", "caveat": "基础记录量不等于全部 AI 功能无限；Granola 明确区分高级摘要限制，不能泛化为没有任何摘要。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C5" }, { "run": "codex", "claim_id": "C6" }, { "run": "claude", "claim_id": "MC7" } ] }, { "product": "Fireflies", "best_for": "小型客户团队愿意配置会议记录到 CRM 的工作流。", "not_for": "要求零维护，或默认免费层就覆盖五人 CRM 工作流。", "rationale": "Otter 的集成描述与 Granola 的销售工作流描述支持把它列为试用候选。", "caveat": "精确套餐、同步深度和五人费用未确认；无机器人模式的对手描述存在前后不一致。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C3" }, { "run": "codex", "claim_id": "C4" }, { "run": "claude", "claim_id": "MC4" }, { "run": "claude", "claim_id": "MC6" } ] }, { "product": "Otter", "best_for": "想单独考察个人转录体验，愿意先检查语言、时长与协作方式。", "not_for": "把此记录理解成“实时协作最佳”或中文识别推荐。", "rationale": "Codex 给了个人转录候选；Claude 认为证据不足，没有给推荐位。保留发现但不升级为共同推荐。", "caveat": "未验证中文表现、实时多人协作或准确率；两轮对免费额度的记录不能证明体验更好。", "agreement_label": "仅 Codex 建议", "disposition": "补充考察 · 证据不足", "evidence": [ { "run": "codex", "claim_id": "C1" }, { "run": "codex", "claim_id": "C2" }, { "run": "claude", "claim_id": "MC1" } ] } ], "differences": [ "Claude 未推荐 Otter；两轮并非三款全部达成推荐共识。", "Claude 未找到 Otter 作为发布者的有效来源，但 Codex 已读取其 Fireflies 对比页，合并后可补这一覆盖缺口。", "来源未提及限制不等于与另一个来源矛盾；Fathom 免费录制与高级摘要额度是不同维度。", "不采用准确率百分比、隐私指控和未实测的中文效果。" ], "next_checks": [ "用同一段经参与者同意的中英混合会议测试说话人、行动项与导出。", "验证五人工作区权限、CRM 同步、存储策略及录制模式。" ] }, "review_notes": [ { "category_id": "meeting-notes", "run": "claude", "claim_id": "MC7", "note": "Granola 原文限制的是 advanced AI summaries；不能据此断言超额会议没有任何 AI 摘要。" } ] }, { "id": "online-forms", "title": "在线表单", "question": "Best online forms for small-team lead capture, conversational surveys and operational workflows?", "candidates": [ "Tally", "Typeform", "Jotform" ], "audience": "3 人小团队，每月约 1000 次提交；区分线索收集、品牌问卷和业务表单。", "scenarios": [ "简单线索收集与条件逻辑", "逐题展示的品牌问卷", "带文件、支付或审批的业务表单" ], "criteria": [ "编辑与填写体验", "提交量限制", "条件逻辑和文件", "支付或审批支持", "品牌与团队功能" ], "special_prompt": "prompts/categories/online-forms.md", "runs": [ { "id": "codex", "metadata": { "run_id": "codex-2026-09-28", "researcher": "Codex", "model_id": "gpt-6-astra", "model_note": "研究与整合模型：GPT-6 Astra，推理强度 high。", "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:35:14.962471+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "execution": "Primary assistant live web research; manually structured findings. No Claude findings read before this run was saved. Category seeds were chosen by this researcher, so discovery independence is limited.", "hands_on_testing": false, "model_display_name": "GPT-6 Astra", "reasoning_effort": "high", "applies_to": [ "codex-2026-09-28", "integration" ] }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "online-forms", "question": "Best online forms for small-team lead capture, conversational surveys and operational workflows?", "scope_limitations": [ "仅比较三个预先指定候选，未覆盖整个市场。", "研究使用有利益关系的竞争对手材料；未做产品实测。" ], "sources": [ { "id": "S1", "url": "https://www.jotform.com/blog/tally-form-alternatives/", "title": "7 best Tally form alternatives", "publisher": "Jotform", "publisher_product": "Jotform", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "What Tally does well; Typeform section", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S2", "url": "https://tally.so/help/best-alternatives-to-typeform", "title": "Best Typeform Alternatives", "publisher": "Tally", "publisher_product": "Tally", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "JotForm pros/cons and submission limits", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S3", "url": "https://www.fillout.com/blog/best-ai-form-builder", "title": "The 10 Best AI Form Builders", "publisher": "Fillout", "publisher_product": "Fillout", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Jotform, Typeform and Tally sections", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" } ], "claims": [ { "id": "C1", "product": "Tally", "statement": "Jotform 和 Fillout 都认可 Tally 的文档式编辑及简单表单用途。", "source_ids": [ "S1", "S3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C2", "product": "Tally", "statement": "Jotform 对复杂自动化和团队流程能力持保留态度；Fillout 描述手工配置逻辑的负担。", "source_ids": [ "S1", "S3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "两类不同限制，不能推导为不存在自动化或条件逻辑。" }, { "id": "C3", "product": "Typeform", "statement": "Jotform 和 Fillout 将 Typeform 的逐题式展示与品牌填写体验作为特点。", "source_ids": [ "S1", "S3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C4", "product": "Typeform", "statement": "Fillout 提醒响应量和去品牌存在套餐限制。", "source_ids": [ "S3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "不采用其完成率提升断言；1000 次提交需要核实当期套餐。" }, { "id": "C5", "product": "Jotform", "statement": "Tally 认可 Jotform 的模板范围，Fillout 描述其提交后的工作流能力。", "source_ids": [ "S2", "S3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "模板和工作流是两个命题，并非同一事实的双重验证。" }, { "id": "C6", "product": "Jotform", "statement": "Tally 和 Fillout 都提醒免费额度、功能复杂度及升级成本。", "source_ids": [ "S2", "S3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "未验证付款费用、审批权限和 3 人协作的完整费用。" } ], "recommendations": [ { "product": "Tally", "best_for": "预算敏感、以简单线索收集为主的小团队", "not_for": "需要开箱即用复杂审批和细粒度团队管理", "rationale": "文档式编辑有两家对手描述；1000 次提交仍需检查公平使用条款。", "claim_ids": [ "C1", "C2" ], "evidence_strength": "moderate", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "Typeform", "best_for": "重视逐题品牌问卷体验的团队", "not_for": "目标是最低费用处理所有业务录入", "rationale": "体验定位有交叉支持，但没有完成率优势证据。", "claim_ids": [ "C3", "C4" ], "evidence_strength": "moderate", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "Jotform", "best_for": "需要模板及提交后业务流程的团队", "not_for": "仅收集少量简单字段且不愿配置流程", "rationale": "模板和工作流构成候选理由，实际支付、文件、审批应逐项验证。", "claim_ids": [ "C5", "C6" ], "evidence_strength": "low", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" } ], "conflicts": [ { "topic": "无限与限额的含义", "source_ids": [ "S1", "S3" ], "description": "宽松免费计划不能推出无限团队权限、文件容量或无支付处理费。", "resolution": "保留条件，避免“无限一切”的表述。" } ], "excluded": [ { "url": "https://www.jotform.com/blog/tally-form-alternatives/", "publisher": "Jotform", "target": "Jotform", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://tally.so/help/best-alternatives-to-typeform", "publisher": "Tally", "target": "Tally", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://www.fillout.com/blog/best-ai-form-builder", "publisher": "Fillout", "target": "Fillout", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" } ], "coverage_gaps": [ "没有用 1000 次提交、3 个协作者计算等价报价。", "没有 A/B 测试证明逐题填写提高转化。" ] } }, { "id": "claude", "metadata": { "run_id": "claude-2026-09-28", "researcher": "Claude Personal", "model_id": "claude-opus-5-5", "model_note": "主研究模型由 CLI init/result 元数据确认为 claude-opus-5-5；工具辅助模型记录包含 claude-haiku-4-5-20251001。", "auxiliary_model_ids": [ "claude-haiku-4-5-20251001" ], "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:40:57.839849+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "runner": "claude-personal CLI 2.1.281", "runner_input": "research/inputs/claude-runner.md", "runner_input_sha256": "09715b90efc5c8fbd280f9d9999dfe26744555c78802c3b0b2a4779c15c4e47c", "execution": "Separate CLI process with only WebSearch/WebFetch tools; no access to Codex findings. Same category briefs and protocol; discovery methods and tool providers differ.", "duration_ms": 614931, "tool_calls": 56, "hands_on_testing": false }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "online-forms", "question": "Best online forms for small-team lead capture, conversational surveys and operational workflows?", "scope_limitations": [ "受众是 3 人小团队，每月约 1000 次提交；只比较 Tally、Typeform、Jotform。", "证据全部来自竞品；一问一页的展示方式不代表转化率更高，不引用任何转化或完成率百分比。", "竞品表格里的安全和合规说法一律不采用。", "以后值得加入：Fillout、Paperform、Google Forms。" ], "sources": [ { "id": "F1", "url": "https://tally.so/help/compare", "title": "Compare Tally: Free Form Builder vs Typeform, Jotform & More", "publisher": "Tally", "publisher_product": "Tally", "relationship": "Typeform 和 Jotform 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "免费档限制表、功能所需档位价格表", "note": "Tally 对自身“不限提交”的说法和 FAQ 属于自我推广，已排除。" }, { "id": "F2", "url": "https://www.jotform.com/jotform-vs-tally-forms/", "title": "Tally Forms vs Jotform | Better Forms, One Plan", "publisher": "Jotform", "publisher_product": "Jotform", "relationship": "Tally 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "功能对比表（模板、支付网关、导出、条件逻辑）", "note": "合规列（PCI/HIPAA/SOC 2）已排除。" }, { "id": "F3", "url": "https://www.jotform.com/blog/tally-form-alternatives/", "title": "7 best Tally form alternatives for 2026", "publisher": "Jotform", "publisher_product": "Jotform", "relationship": "Tally 和 Typeform 的直接竞争对手（与 F2 同属 Jotform 家族）", "published_at": "2026-09-14", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Tally 段、Typeform 段、功能对比", "note": "作者 Harry Guinness；“Why Jotform is more than…”一段属于自我推广，已排除。" }, { "id": "F4", "url": "https://www.typeform.com/blog/product/typeform-vs-jotform/", "title": "Typeform vs. Jotform: Best Form Builder? [2026]", "publisher": "Typeform", "publisher_product": "Typeform", "relationship": "Jotform 的直接竞争对手", "published_at": "2026-09-04", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Jotform 让步清单、限制清单、Jotform 价格表", "note": "页面显示更新日期 2026-09-03、发布日期 2026-09-04；“87% 用户更高回复率”已排除。" }, { "id": "F5", "url": "https://www.typeform.com/blog/typeform-vs-tally", "title": "Typeform vs. Tally: Which is better? [2025]", "publisher": "Typeform", "publisher_product": "Typeform", "relationship": "Tally 的直接竞争对手", "published_at": "2025-07-03", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Tally 优点、分析缺口、集成限制", "note": "页面显示的更新日期（2025-05-23）早于发布日期，日期信息不一致；内容可能偏旧。" }, { "id": "F6", "url": "https://tally.so/help/best-alternatives-to-typeform", "title": "Best Typeform Alternatives: Comparison in 2026", "publisher": "Tally", "publisher_product": "Tally", "relationship": "Typeform 和 Jotform 的直接竞争对手（与 F1 同属 Tally 家族）", "published_at": "2026-09-25", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Typeform 段、JotForm 段", "note": "全文在推广 Tally；只采用对他方的描述。" } ], "claims": [ { "id": "FC1", "product": "Tally", "statement": "Jotform 和 Typeform 都承认 Tally 的免费档很慷慨，编辑器像写文档，适合快速做标准表单（Typeform：类似 Notion；Jotform：快速易用）。", "source_ids": [ "F3", "F5" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "“不限提交”是 Tally 自己的说法，已排除；每月 1000 次提交是否适用公平使用条款没有核实。" }, { "id": "FC2", "product": "Tally", "statement": "Jotform 称 Tally 的自动化、深度集成和团队协作较基础，支付网关只有 Stripe，导出只有 CSV；Typeform 称 Tally 免费档只有 6 个直接集成，缺内建的流失率等深度分析。", "source_ids": [ "F2", "F3", "F5" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "集成和协作偏浅有两家印证；Stripe 独家、只能导出 CSV 只有 Jotform 一家来源。F5 可能已过时。" }, { "id": "FC3", "product": "Tally", "statement": "Jotform 的两个页面口径不一：F2 说 Tally 有和 Jotform 一样的条件逻辑，F3 说 Tally 的条件逻辑“有限”。", "source_ids": [ "F2", "F3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "同一家族内自相矛盾，见冲突项。" }, { "id": "FC4", "product": "Typeform", "statement": "Tally 和 Jotform 都承认 Typeform 是一问一页对话式表单的代表，体验精致、有高级感，条件逻辑和集成较强。", "source_ids": [ "F6", "F3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "展示方式不等于经过验证的转化提升。" }, { "id": "FC5", "product": "Typeform", "statement": "三个来源都指向成本和额度问题：Tally 称免费档每月只有 10 次回复，文件上传和支付要 $39+/月档，去品牌要 $79/月；Jotform 说它“可能很贵”；Tally 说团队常在低价档碰到提交上限。", "source_ids": [ "F1", "F3", "F6" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "F1 和 F6 同属 Tally 家族；每月 1000 次回复要哪一档不清楚。" }, { "id": "FC6", "product": "Typeform", "statement": "Jotform 和 Tally 都说一问一页的形式本身灵活性较低。", "source_ids": [ "F3", "F6" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "这是利益相关方的判断，适不适合要看具体问卷。" }, { "id": "FC7", "product": "Jotform", "statement": "Tally 承认 Jotform 模板库很大、支持 40+ 支付处理器、有审批工作流；Typeform 承认 Jotform 支持收款、模板、答案引用，并有在线支持团队。", "source_ids": [ "F6", "F4" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "审批工作流只有 Tally 一家提到；能收款不代表没有处理手续费。" }, { "id": "FC8", "product": "Jotform", "statement": "Tally 称 Jotform 免费档限 5 个表单、每月 100 次提交，界面可能让人觉得繁杂，涨价主要是提高额度而不是加功能；Typeform 称 Bronze 档 25 个表单，不限表单要 Enterprise。", "source_ids": [ "F1", "F6", "F4" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "免费档限制有 Tally 家族两页一致的说法；每月 1000 次提交需要付费档，但 Bronze 档的提交额度没有竞品给出。" }, { "id": "FC9", "product": "Jotform", "statement": "Typeform 称 Jotform 缺少逐题流失分析，不建议嵌入邮件，自定义域名只在 Enterprise 档。", "source_ids": [ "F4" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家来源；自定义域名的档位和 Tally 的说法冲突。" } ], "recommendations": [ { "product": "Tally", "best_for": "场景一：简单线索收集加条件逻辑，想先用免费档或低价档的小团队", "not_for": "需要多种支付网关、审批流或深度内建分析的业务表单", "rationale": "两个竞品都承认它的免费档和编辑体验（FC1）；局限在集成、协作和支付网关（FC2）。", "claim_ids": [ "FC1", "FC2", "FC3" ], "evidence_strength": "moderate", "caveat": "每月 1000 次提交在免费档是否受公平使用限制没有核实；条件逻辑深度说法有冲突。" }, { "product": "Typeform", "best_for": "场景二：重视品牌感、逐题展示的问卷", "not_for": "预算敏感、需要每月约 1000 次回复但不想上高价档的团队；需要灵活多题同页布局的场景", "rationale": "两个竞品都承认它在对话式和精致体验上的定位（FC4）；三个来源都提示成本和额度问题（FC5）。", "claim_ids": [ "FC4", "FC5", "FC6" ], "evidence_strength": "low", "caveat": "每月 1000 次回复对应哪一档、多少钱都不清楚；没有转化提升的证据。" }, { "product": "Jotform", "best_for": "场景三：带文件上传、收款、审批的运营类业务表单", "not_for": "只要极简表单、想长期留在免费档的团队（免费档每月 100 次提交）", "rationale": "Tally 和 Typeform 都承认它的支付和模板能力（FC7）；免费档额度有限、需要付费（FC8）。", "claim_ids": [ "FC7", "FC8", "FC9" ], "evidence_strength": "moderate", "caveat": "Bronze 档价格有冲突，提交额度没有核实；审批只有一家来源；支付处理费另计。" } ], "conflicts": [ { "topic": "Jotform Bronze 档价格", "source_ids": [ "F1", "F4", "F6" ], "description": "Tally 比较页称 $34+/月，Typeform 称 $34/月，Tally 的 Typeform 替代页称 $39/月起；都没有说明是年付还是月付。", "resolution": "未解决；不做价格排序。" }, { "topic": "Typeform 起价", "source_ids": [ "F3", "F1" ], "description": "Jotform 称 Typeform 从 $29/月起；Tally 称文件上传和支付要 $39+/月档。", "resolution": "可能是不同档位对应不同功能，没法对齐；未解决。" }, { "topic": "Jotform 自定义域名档位", "source_ids": [ "F1", "F4" ], "description": "Tally 称 $39+/月起；Typeform 称只在 Enterprise 档。", "resolution": "未解决。" }, { "topic": "Tally 条件逻辑深度", "source_ids": [ "F2", "F3" ], "description": "Jotform 的两个页面分别说“与 Jotform 相同”和“有限”。", "resolution": "未解决；建议试用时验证具体分支需求。" } ], "excluded": [ { "url": "https://tally.so/help/compare", "publisher": "Tally", "target": "Tally", "reason": "Tally 自称免费不限表单和提交、Pro $29/月、FAQ 里“Tally 更好”等属于自我推广。" }, { "url": "https://www.jotform.com/jotform-vs-tally-forms/", "publisher": "Jotform", "target": "Tally", "reason": "关于 Tally 缺 PCI/HIPAA/SOC 2 的合规说法，按协议不能只凭竞品表格采用。" }, { "url": "https://www.typeform.com/blog/typeform-vs-tally", "publisher": "Typeform", "target": "Tally", "reason": "“Tally 不符合 HIPAA”属于合规断言，不采用。" }, { "url": "https://www.typeform.com/blog/product/typeform-vs-jotform/", "publisher": "Typeform", "target": "Typeform", "reason": "“87% 用户获得更高回复率”等转化或营销数字属于自我推广。" }, { "url": "https://stackcoast.com/typeform-vs-jotform-vs-tally/", "publisher": "StackCoast", "target": "三者", "reason": "中立或第三方评测，只用于发现线索（包括“Jotform $34/月 1000 次回复”这类数字）。" } ], "coverage_gaps": [ "三者支撑每月约 1000 次提交所需的档位和实际价格，都没有经过竞品之外的核实。", "文件上传的大小和存储上限、支付手续费，竞品来源基本都没有提到。", "Typeform 对 Tally 的对比页（F5）是 2025 年的，可能已过时。", "Tally 的公平使用条款对“不限提交”的约束没有证据。", "团队席位和协作功能（3 人团队）在各产品哪一档提供，证据不足。" ] } } ], "integration": { "category_id": "online-forms", "summary": "三款在简单收集、逐题品牌体验和提交后业务流程上各有试用理由；不把厂商的“转化更高”宣传变成研究结论。", "entries": [ { "product": "Tally", "best_for": "三人小团队快速制作文档式线索收集表单。", "not_for": "未经验证就要求复杂审批、精细角色权限或特定支付渠道。", "rationale": "两轮均从 Jotform 等对手材料提取到简单编辑的定位。", "caveat": "“支持条件逻辑”与“逻辑复杂度有限”并非直接矛盾；不能声称它与 Jotform 逻辑能力相同。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C1" }, { "run": "codex", "claim_id": "C2" }, { "run": "claude", "claim_id": "FC1" } ] }, { "product": "Typeform", "best_for": "重视逐题式填写和品牌展示，愿意核对响应量预算。", "not_for": "长表单需要一页浏览，或预算不允许提交量和去品牌升级。", "rationale": "两轮对逐题体验定位一致；这种设计是否适合你的受众需实测。", "caveat": "没有独立完成率测试；每月一千次提交、三名成员的实际档位未确认。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C3" }, { "run": "codex", "claim_id": "C4" }, { "run": "claude", "claim_id": "FC4" } ] }, { "product": "Jotform", "best_for": "表单之后还要处理模板、收款或审批流程的团队。", "not_for": "只要极简收集，且无法投入时间梳理功能与额度。", "rationale": "Codex 的 Tally 与 Fillout 证据支持模板和后续工作流；Claude 也给出相同方向。", "caveat": "支付交易费、提交额度与团队席位未统一报价；不采纳未复核的域名和分析缺失断言。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C5" }, { "run": "codex", "claim_id": "C6" } ] } ], "differences": [ "Claude FC3 将 Jotform 表中“双方都有条件逻辑”解释为“逻辑一样”，超出了来源；整合时纠正，原始输出保留批注。", "两轮有多篇同源 Jotform/Tally 材料，不能按页面数重复增加可信度。", "免费额度、域名档位和价格数字有日期与口径差异，暂不排序成本。" ], "next_checks": [ "三人角色配置下测试一千次月提交、导出、支付和审批的总成本。", "用同一表单内容做可用性测试，区分设计偏好与真实完成率。" ] }, "review_notes": [ { "category_id": "online-forms", "run": "claude", "claim_id": "FC3", "note": "复读 F2 只证实 Tally 有条件逻辑，不证实两款能力相同；与“能力有限”不构成直接矛盾。" } ] }, { "id": "email-marketing", "title": "邮件营销", "question": "Best email marketing for a small business with 2000 opted-in contacts and 4 monthly campaigns?", "candidates": [ "MailerLite", "Brevo", "Mailchimp" ], "audience": "已有 2000 名许可订阅者的小企业，每月 4 次群发（约 8000 封），另行区分交易邮件。", "scenarios": [ "简单 newsletter 与欢迎流程", "较大名单但低发送频率", "需要更多营销集成与分群" ], "criteria": [ "联系人与发送量计费", "欢迎流程与分群", "编辑门槛", "集成与导出", "费用包含项" ], "special_prompt": "prompts/categories/email-marketing.md", "runs": [ { "id": "codex", "metadata": { "run_id": "codex-2026-09-28", "researcher": "Codex", "model_id": "gpt-6-astra", "model_note": "研究与整合模型：GPT-6 Astra，推理强度 high。", "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:35:14.962471+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "execution": "Primary assistant live web research; manually structured findings. No Claude findings read before this run was saved. Category seeds were chosen by this researcher, so discovery independence is limited.", "hands_on_testing": false, "model_display_name": "GPT-6 Astra", "reasoning_effort": "high", "applies_to": [ "codex-2026-09-28", "integration" ] }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "email-marketing", "question": "Best email marketing for a small business with 2000 opted-in contacts and 4 monthly campaigns?", "scope_limitations": [ "仅比较三个预先指定候选，未覆盖整个市场。", "研究使用有利益关系的竞争对手材料；未做产品实测。" ], "sources": [ { "id": "S1", "url": "https://www.mailerlite.com/blog/mailchimp-alternatives", "title": "10 Mailchimp Alternatives", "publisher": "MailerLite", "publisher_product": "MailerLite", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Why use an alternative; Brevo for infrequent senders", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S2", "url": "https://moosend.com/blog/mailerlite-alternatives/", "title": "12 Best MailerLite Alternatives", "publisher": "Moosend", "publisher_product": "Moosend", "relationship": "competing_vendor", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "MailerLite limitations; Mailchimp section; March 2026 information disclaimer", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" }, { "id": "S3", "url": "https://www.beehiiv.com/blog/brevo-alternatives", "title": "15 Best Brevo Alternatives", "publisher": "beehiiv", "publisher_product": "beehiiv", "relationship": "competing_vendor", "published_at": "2026-05-05", "accessed_at": "2026-09-28", "access_status": "read", "locator": "Brevo introduction; Mailchimp; MailerLite", "note": "对手材料有商业动机；只纳入评价其他产品的段落。" } ], "claims": [ { "id": "C1", "product": "MailerLite", "statement": "Moosend 与 beehiiv 将 MailerLite 描述为较易上手的小企业邮件工具。", "source_ids": [ "S2", "S3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C2", "product": "MailerLite", "statement": "Moosend 提醒高级自动化、报表和分群边界；beehiiv 区分邮件工具与 CRM、多渠道能力。", "source_ids": [ "S2", "S3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "各命题来自不同来源；不推导为没有基础欢迎流程。" }, { "id": "C3", "product": "Brevo", "statement": "MailerLite 将 Brevo 视为大名单低频发送的候选；beehiiv 描述按发送量计费的取舍。", "source_ids": [ "S1", "S3" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "竞争对手说法，未实测或独立核实。" }, { "id": "C4", "product": "Brevo", "statement": "beehiiv 提醒免费日发送限制和多渠道界面的学习负担。", "source_ids": [ "S3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "未独立验证当前上限；每月可发不等于单日能群发全部联系人。" }, { "id": "C5", "product": "Mailchimp", "statement": "Moosend 描述 Mailchimp 的 A/B 测试能力，beehiiv 描述其营销模板和集成范围。", "source_ids": [ "S2", "S3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "不表示两个来源验证了相同功能，也不保证低价套餐包含。" }, { "id": "C6", "product": "Mailchimp", "statement": "MailerLite 与 beehiiv 都认为免费计划有限，但给出的发送额度不同。", "source_ids": [ "S1", "S3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "当前免费额度不确定，不能拼接不同版本数据。" } ], "recommendations": [ { "product": "MailerLite", "best_for": "重视简单 newsletter 和基础欢迎流程的小企业", "not_for": "需要复杂跨渠道与深度 CRM 的团队", "rationale": "易用定位有两家对手支持；欢迎流程触发器仍需按套餐核实。", "claim_ids": [ "C1", "C2" ], "evidence_strength": "moderate", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "Brevo", "best_for": "名单较大而发送频率较低的使用者", "not_for": "要求仅凭免费计划一次发完整大名单", "rationale": "发送量计费值得按 8000 封/月试算，但本轮不宣称最便宜。", "claim_ids": [ "C3", "C4" ], "evidence_strength": "moderate", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" }, { "product": "Mailchimp", "best_for": "需要广泛营销集成并愿意核实功能套餐的团队", "not_for": "仅凭免费额度或起价决定 2000 人名单方案", "rationale": "集成与测试能力形成试用理由；完整费用及限制尚未核实。", "claim_ids": [ "C5", "C6" ], "evidence_strength": "low", "caveat": "试用候选，不是已验证赢家；采购前复核当前套餐与实际工作流。" } ], "conflicts": [ { "topic": "Mailchimp 免费发送额度", "source_ids": [ "S1", "S3" ], "description": "MailerLite 页面写 1000 封/月，beehiiv 表格写 500 封/月；时间与版本未统一。", "resolution": "不选一个数字冒充现价；两者都不足以覆盖本场景 8000 封/月。" } ], "excluded": [ { "url": "https://www.brevo.com/blog/best-email-marketing-services/", "publisher": "Brevo", "target": "MailerLite / Mailchimp", "reason": "读取返回无可用正文；搜索摘要不算已经检查的证据。" }, { "url": "https://www.mailerlite.com/blog/mailchimp-alternatives", "publisher": "MailerLite", "target": "MailerLite", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://moosend.com/blog/mailerlite-alternatives/", "publisher": "Moosend", "target": "Moosend", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" }, { "url": "https://www.beehiiv.com/blog/brevo-alternatives", "publisher": "beehiiv", "target": "beehiiv", "reason": "该页对出版方自身产品的推荐、购买引导及自荐评分不用于推荐证据。" } ], "coverage_gaps": [ "8000 封仅为四次群发，不包含欢迎流程和重试。", "未确认 2000 联系人的当前等价套餐报价。", "没有独立送达率或收件箱投递测试。" ] } }, { "id": "claude", "metadata": { "run_id": "claude-2026-09-28", "researcher": "Claude Personal", "model_id": "claude-opus-5-5", "model_note": "主研究模型由 CLI init/result 元数据确认为 claude-opus-5-5；工具辅助模型记录包含 claude-haiku-4-5-20251001。", "auxiliary_model_ids": [ "claude-haiku-4-5-20251001" ], "cutoff": "2026-09-28", "completed_at": "2026-09-27T18:40:57.839849+00:00", "prompt_version": "1.0.0", "input_hashes": { "workflow-automation": "01cb5f18a4ed94b16aa6b6b5c9524fec590c369201eb8717e5d9c8ba37048723", "project-management": "2dc42455833f42d5748d5a03eb9fed5c6d414dfa00e77eb5108a3c874d3f0b8d", "meeting-notes": "e1b084aa746f2668d77019a1711f34da3519deea551c155b8647a44952c78e2c", "online-forms": "17b4f1fb4374ff05fd675446b5ba4b1e1cda135906190ff4f12c4c8a84df4e8c", "email-marketing": "48eec9b2337e5edab21dab0a1ab741489fe78f17d61d30cb6d06ab4ae77a021c" }, "runner": "claude-personal CLI 2.1.281", "runner_input": "research/inputs/claude-runner.md", "runner_input_sha256": "09715b90efc5c8fbd280f9d9999dfe26744555c78802c3b0b2a4779c15c4e47c", "execution": "Separate CLI process with only WebSearch/WebFetch tools; no access to Codex findings. Same category briefs and protocol; discovery methods and tool providers differ.", "duration_ms": 614931, "tool_calls": 56, "hands_on_testing": false }, "review_status": "original-output-with-editorial-notes; not-verified-facts", "research": { "category_id": "email-marketing", "question": "Best email marketing for a small business with 2000 opted-in contacts and 4 monthly campaigns?", "scope_limitations": [ "受众是有 2000 名许可订阅者、每月约 8000 封群发（不含自动化）的小企业；交易邮件单独考虑。只比较 MailerLite、Brevo、Mailchimp。", "Brevo 发布的 4 个对比页都没有读取到（页面内容无法获取），所以 MailerLite 和 Mailchimp 缺少来自 Brevo 的视角。", "不从竞品营销推断送达率或进箱率，也不采纳任何法律合规说法。", "以后值得加入：Kit、EmailOctopus、Klaviyo（电商场景）。" ], "sources": [ { "id": "E1", "url": "https://www.mailerlite.com/mailchimp-alternative", "title": "MailerLite vs. Mailchimp: Free Mailchimp Alternative", "publisher": "MailerLite", "publisher_product": "MailerLite", "relationship": "Mailchimp 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "计费联系人说明、按订阅者数的价格表、发送上限", "note": "MailerLite 自身价格和满意度数字已排除。" }, { "id": "E2", "url": "https://www.mailerlite.com/brevo-alternative", "title": "Free Brevo (Formerly Sendinblue) Alternative", "publisher": "MailerLite", "publisher_product": "MailerLite", "relationship": "Brevo 的直接竞争对手（与 E1 同属 MailerLite 家族）", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "Brevo 定价模型、功能限制清单", "note": "提到 2025 年底 Brevo 调整了定价；关于送达率的用户证言已排除。" }, { "id": "E3", "url": "https://mailchimp.com/resources/mailchimp-vs-mailerlite/", "title": "Comparing Mailchimp vs Mailerlite", "publisher": "Mailchimp", "publisher_product": "Mailchimp", "relationship": "MailerLite 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "定价让步、自动化、分析、集成、团队席位", "note": "ROI 和用户规模等数字已排除。" }, { "id": "E4", "url": "https://mailchimp.com/resources/mailchimp-vs-brevo/", "title": "Comparing Brevo vs Mailchimp", "publisher": "Mailchimp", "publisher_product": "Mailchimp", "relationship": "Brevo 的直接竞争对手（与 E3 同属 Mailchimp 家族）", "published_at": null, "accessed_at": "2026-09-28", "access_status": "read", "locator": "定价模型让步、多渠道、席位和支持限制、Brevo 价格", "note": "“#1 平台”、27x ROI 等已排除。" }, { "id": "E5", "url": "https://www.brevo.com/mailerlite/", "title": "Brevo vs. MailerLite: Get Low Pricing For Email and More", "publisher": "Brevo", "publisher_product": "Brevo", "relationship": "MailerLite 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "inaccessible", "locator": "只拿到标题", "note": "尝试了两次都没拿到正文，没有引用。" }, { "id": "E6", "url": "https://www.brevo.com/mailchimp/", "title": "Mailchimp vs. Brevo: 9,000 emails free/month", "publisher": "Brevo", "publisher_product": "Brevo", "relationship": "Mailchimp 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "inaccessible", "locator": "只拿到标题", "note": "没有引用。" }, { "id": "E7", "url": "https://www.brevo.com/blog/mailchimp-prices/", "title": "Mailchimp Prices by Plan, Plus Add-ons & Extra Costs", "publisher": "Brevo", "publisher_product": "Brevo", "relationship": "Mailchimp 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "inaccessible", "locator": "只拿到标题", "note": "没有引用。" }, { "id": "E8", "url": "https://www.brevo.com/blog/cheap-email-marketing/", "title": "18 Cheapest Email Marketing Platforms Compared (2026)", "publisher": "Brevo", "publisher_product": "Brevo", "relationship": "MailerLite 和 Mailchimp 的直接竞争对手", "published_at": null, "accessed_at": "2026-09-28", "access_status": "inaccessible", "locator": "只拿到标题", "note": "没有引用。" } ], "claims": [ { "id": "EC1", "product": "MailerLite", "statement": "Mailchimp 承认 MailerLite 入门价格更低（付费档从 $9/月起），免费档覆盖 500 名订阅者（Mailchimp 是 250 个联系人），定位适合基础邮件和创作者。", "source_ids": [ "E3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有 Mailchimp 一家竞品来源；价格的计费周期没有说明。" }, { "id": "EC2", "product": "MailerLite", "statement": "Mailchimp 称 MailerLite 的自动化上限是 100 步（Mailchimp 是 200 步），没有收入归因和行业基准分析，集成约 100+，“Growing Business”档只有 3 个用户席位，没有电话支持。", "source_ids": [ "E3" ], "kind": "competitor_claim", "status": "unverified", "caveat": "100 步对欢迎流程绰绰有余；这些限制主要影响多分支旅程和归因分析。档位名称和 E1 不一致。" }, { "id": "EC3", "product": "Brevo", "statement": "Mailchimp 和 MailerLite 都说 Brevo 主要按发送量计费，而不是按联系人数；Mailchimp 承认 Brevo 在同一平台里提供交易邮件、SMS、WhatsApp、CRM 等多渠道能力，入门档更便宜（Starter $8/月）。", "source_ids": [ "E4", "E2" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "MailerLite 同时说 Brevo 在 2025 年底调整后加了联系人上限，这对 2000 名联系人的计费有直接影响，见冲突项。" }, { "id": "EC4", "product": "Brevo", "statement": "MailerLite 称 Brevo 把落地页移出了 Starter，A/B 测试和自动化按档位限制，去品牌要约 $12/月加购；Mailchimp 称 Brevo 的 Standard 档默认只有 1 个用户席位，电话支持只在 Professional 档（约 $449/月）及以上。", "source_ids": [ "E2", "E4" ], "kind": "competitor_claim", "status": "corroborated_by_competitors", "caveat": "“功能按档限制”有两家印证；具体数字都只有一家来源。" }, { "id": "EC5", "product": "Brevo", "statement": "MailerLite 称 Brevo 免费档每天 300 封，还列出了按月发送量分级的价格（5K 封 $18、10K 封 $35 等），并称基础档有“500 个联系人、$9”的上限。", "source_ids": [ "E2" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家来源。每日 300 封的上限对每月 4 次、每次 2000 封的群发不够用（推断）；价格没有说明币种和周期，不计算总额。" }, { "id": "EC6", "product": "Mailchimp", "statement": "MailerLite 称 Mailchimp 按联系人计费，而且会把已订阅、已退订和非订阅的联系人都算进去；每月发送上限是联系人总数的 12 倍。", "source_ids": [ "E1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家来源（Brevo 的对比页没读到）；退订联系人怎么计费，对 2000 人名单的成本影响很大，需要自行核实。" }, { "id": "EC7", "product": "Mailchimp", "statement": "MailerLite 表格称 Mailchimp Standard 档 2,500 名联系人以内是 $60/月，5,000 名以内是 $100/月，并说 Mailchimp 没有年付折扣。", "source_ids": [ "E1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "只有一家来源；没有说明取价日期；档位功能不对等，不能直接比较成本。" }, { "id": "EC8", "product": "Mailchimp", "statement": "MailerLite 承认 Mailchimp 也有拖拽编辑器和模板库，付费档提供 24/7 邮件和在线聊天支持。", "source_ids": [ "E1" ], "kind": "competitor_claim", "status": "unverified", "caveat": "这是弱让步，而且只有一家来源；Mailchimp 在营销集成和分群上的优势只有它自己的说法（已排除）。" } ], "recommendations": [ { "product": "MailerLite", "best_for": "场景一：简单 newsletter 加欢迎流程，2000 名订阅者、每月 4 次群发", "not_for": "需要收入归因、复杂多分支旅程或超过 3 个协作席位的团队", "rationale": "Mailchimp 承认它价格更低、免费档更大（EC1）；它被指出的限制主要在高级自动化和分析（EC2），对欢迎流程影响不大（推断）。", "claim_ids": [ "EC1", "EC2" ], "evidence_strength": "low", "caveat": "只有 Mailchimp 一个竞品家族的证据，没达到两个竞品发布者的要求（Brevo 页面没读到）。2000 名订阅者的实际档位价格没有核实。" }, { "product": "Brevo", "best_for": "场景二：名单较大但发送频率低，并希望在同一平台处理交易邮件的团队", "not_for": "需要多人协作席位、又不想上高价档的团队；依赖落地页或 A/B 测试但只买低档的团队", "rationale": "两个竞品都说它按发送量计费（EC3），也都指出它按档限制功能和席位（EC4）。", "claim_ids": [ "EC3", "EC4", "EC5" ], "evidence_strength": "low", "caveat": "2025 年底加入的联系人上限可能让 2000 名联系人需要更高档，冲突没解决；免费档每日 300 封不够一次群发 2000 封（推断）。" } ], "conflicts": [ { "topic": "Brevo 是否对联系人数设上限或收费", "source_ids": [ "E2", "E4" ], "description": "MailerLite 称 Brevo 在 2025 年底调整后加了联系人上限（基础档 500 个联系人、$9）；Mailchimp 只描述它按发送量计费，没提联系人上限。", "resolution": "未解决；这对 2000 名联系人的成本至关重要，要到现行价目上核实。" }, { "topic": "Brevo 入门价", "source_ids": [ "E4", "E2" ], "description": "Mailchimp 称 Starter $8/月；MailerLite 称基础价 $9。", "resolution": "未解决；可能是年付和月付的差别，周期没有说明。" }, { "topic": "MailerLite 档位名称和团队席位", "source_ids": [ "E3", "E1" ], "description": "Mailchimp 引用“Growing Business”档（3 个席位）；MailerLite 的对比表用的是“Power”档。", "resolution": "可能是套餐改版造成的，未解决。" } ], "excluded": [ { "url": "https://www.mailerlite.com/brevo-alternative", "publisher": "MailerLite", "target": "Brevo", "reason": "用户证言称 Brevo 有送达率问题，按协议不从竞品营销推断送达率。" }, { "url": "https://www.mailerlite.com/mailchimp-alternative", "publisher": "MailerLite", "target": "MailerLite", "reason": "MailerLite 自身价格（Power 档 $25–$389）、97% 满意度、5 分钟响应等属于自我推广。" }, { "url": "https://mailchimp.com/resources/mailchimp-vs-mailerlite/", "publisher": "Mailchimp", "target": "Mailchimp", "reason": "9x 和 27x ROI、200 步自动化、330+ 集成等 Mailchimp 自身描述不能支撑它的推荐。" }, { "url": "https://mailchimp.com/resources/mailchimp-vs-brevo/", "publisher": "Mailchimp", "target": "Mailchimp", "reason": "“#1 平台”、97% 点击率提升、G2 Grid Leader 等属于自我推广。" }, { "url": "https://www.emailtooltester.com/en/blog/brevo-vs-mailchimp/", "publisher": "EmailTooltester", "target": "Brevo、Mailchimp", "reason": "中立或联盟评测站，与候选的关系不明，只用于发现线索。" } ], "coverage_gaps": [ "Brevo 的 4 个对比页都读取失败：MailerLite 和 Mailchimp 各自只有一个竞品发布者的证据，没达到“每个候选至少两个竞品发布者”的要求。", "场景三（更多营销集成和分群）不给推荐：Mailchimp 在这方面的优势只有它自己的说法，竞品只给了很弱的让步。", "2000 名联系人、每月 8000 封在三家各需要哪一档、多少钱，都没有从现行价目核实，不能比较费用。", "退订和重复联系人怎么计费，只有 MailerLite 对 Mailchimp 的一面之词。", "交易邮件：只有 Brevo 被竞品确认包含交易邮件能力；MailerLite 和 Mailchimp 的交易邮件（例如 Mandrill）费用没有证据。", "每日发送上限、导出能力这两项评估标准，竞品来源覆盖不足。" ] } } ], "integration": { "category_id": "email-marketing", "summary": "按 2,000 联系人、每月四次全量发送（8,000 封，自动化另计）评估。当前证据不足以选出最便宜的一款。", "entries": [ { "product": "MailerLite", "best_for": "以常规 newsletter 和基础自动化为主的小团队。", "not_for": "需要复杂收入归因、跨渠道 CRM 或大量细粒度角色权限。", "rationale": "Codex 的 Moosend、beehiiv 与 Claude 的 Mailchimp 材料都指向基础邮件用途。", "caveat": "Mailchimp 的比较有明确的 2025 年价格基准；其自身数字已从整合依据中排除，不保证当前低价。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C1" }, { "run": "codex", "claim_id": "C2" }, { "run": "claude", "claim_id": "EC1" }, { "run": "claude", "claim_id": "EC2" } ] }, { "product": "Brevo", "best_for": "名单相对较大但发送频率有限，愿意核对联系人与发送量双重限制。", "not_for": "默认联系人无限，或直接用最低入门价覆盖 2,000 联系人和自动化。", "rationale": "两轮都看到按发送量计费的定位；Claude 又找到联系人上限提示，使候选条件更严格。", "caveat": "MailerLite 页面列有基础档联系人限制，与单纯“只按发送量”理解不一致。未核实具体套餐前不能算本场景总价。", "agreement_label": "两轮场景一致", "disposition": "条件候选 · 未实测", "evidence": [ { "run": "codex", "claim_id": "C3" }, { "run": "codex", "claim_id": "C4" }, { "run": "claude", "claim_id": "EC5" } ] }, { "product": "Mailchimp", "best_for": "已有营销集成需求，愿意将 A/B 测试与工作流作为试用重点。", "not_for": "仅因品牌知名度就接受价格，或需要本研究已经确认的成本优势。", "rationale": "Codex 从 Moosend 与 beehiiv 提取到测试和集成用途；Claude 正向对手材料较薄，没有推荐它。", "caveat": "保留单轮建议，不宣称两轮共识。免费额度口径相互冲突，也没有独立送达率比较。", "agreement_label": "仅 Codex 建议", "disposition": "补充考察 · 证据不足", "evidence": [ { "run": "codex", "claim_id": "C5" }, { "run": "codex", "claim_id": "C6" } ] } ], "differences": [ "Claude 未推荐 Mailchimp；合并时保留 Codex 的单轮发现并降低呈现力度。", "Brevo 的联系人限制直接影响 2,000 人场景；两轮的按发送量定位不构成无限联系人保证。", "Claude 的 MailerLite 命题混入了 Mailchimp 自身额度；这些自述数字已排除于最终推荐依据。", "Brevo 发布的四个比较页面读取失败，不能把未读取的内容当成证据。" ], "next_checks": [ "核对 2,000 联系人、8,000 基础发送量、自动化额外发送、退订联系人计数、席位及税费。", "在同一已获许可的名单样本中测试模板、退订、迁移和自动化，不依据厂商自述排送达率。" ] }, "review_notes": [ { "category_id": "email-marketing", "run": "claude", "claim_id": "EC1", "note": "括号中 Mailchimp 自身免费联系人数字是自述，排除；只保留其对 MailerLite 的描述。价格基准为 2025 年。" }, { "category_id": "email-marketing", "run": "claude", "claim_id": "EC2", "note": "括号中 Mailchimp 自身自动化步数是自述，排除；本条未进行功能实测。" } ] } ] }