Expand Chinese candidates and audit data-vendor revenue evidence

apps/docs/scripts/auth-cli.ts

@@ -182,10 +182,10 @@ async function main() {
   assert.equal(researchResponse.status, 200, "authenticated research API");
   assert.ok(researchResponse.headers.get("cache-control")?.includes("no-store"));
   const research = await researchResponse.json();
-  assert.equal(research.companies.length, 1026);
+  assert.equal(research.companies.length, 1028);
   assert.equal(research.categories.length, 11);
-  assert.equal(research.evidence.length, 1374);
-  assert.equal(research.observations.length, 190);
+  assert.equal(research.evidence.length, 1392);
+  assert.equal(research.observations.length, 191);
   assert.equal((await request("/api/research/ai-companies", { method: "POST", headers })).status, 405);
   const researchPage = await request("/docs/other-research/ai-company-research/overview", { headers });
   assert.equal(researchPage.status, 200);

apps/docs/tests/research.test.ts

@@ -23,9 +23,9 @@ test("research revisions share the authentication database and preserve replay l
     for (let i = 0; i < 20; i++) consumeCode(config, "invalid");
     assert.equal(consumeCode(config, "invalid"), "limited");
     const initial = readResearchSnapshot();
-    assert.equal(initial.companies.length, 1026);
-    assert.equal(initial.evidence.length, 1374);
-    assert.equal(initial.observations.length, 190);
+    assert.equal(initial.companies.length, 1028);
+    assert.equal(initial.evidence.length, 1392);
+    assert.equal(initial.observations.length, 191);
     assert.deepEqual(initial.candidateBatches, seed.candidateBatches, "candidate gaps persist in the same database metadata");
     assert.deepEqual(initial.directoryRechecks, seed.directoryRechecks, "follow-up classifications persist without rewriting the discovery audit");
     assert.deepEqual(readResearchSnapshot(), initial, "reopening does not duplicate records");
@@ -64,9 +64,9 @@ test("research validation rejects broken evidence and unsafe source links", () =
 
 test("reviews link explicit companies without turning global or excluded cases into companies", () => {
   const snapshot = structuredClone(seed) as ResearchSnapshot;
-  assert.equal(snapshot.reviews.filter((r) => r.scope === "companies").length, 68);
+  assert.equal(snapshot.reviews.filter((r) => r.scope === "companies").length, 73);
   assert.equal(snapshot.reviews.filter((r) => r.scope === "dataset").length, 10);
-  assert.equal(snapshot.reviews.filter((r) => r.scope === "excluded-candidate").length, 34);
+  assert.equal(snapshot.reviews.filter((r) => r.scope === "excluded-candidate").length, 35);
   const review = snapshot.reviews.find((r) => r.id === "R4")!;
   assert.deepEqual(reviewCompanyIds(snapshot, review), ["AI007", "AI013", "AI021", "AI034"]);
   assert.deepEqual(reviewCompanyIds(snapshot, snapshot.reviews.find((r) => r.id === "CN-R-PSI")!), []);
@@ -85,11 +85,11 @@ test("reviews link explicit companies without turning global or excluded cases i
 test("disclosure counts exclude client income, biographies, targets and duplicate milestones", () => {
   const snapshot = seed as ResearchSnapshot;
   const summary = categorySummary(snapshot);
-  assert.equal(summary.reduce((n, c) => n + c.count, 0), 1026);
+  assert.equal(summary.reduce((n, c) => n + c.count, 0), 1028);
   assert.equal(summary.reduce((n, c) => n + c.self, 0), 27);
-  assert.equal(summary.reduce((n, c) => n + c.estimate, 0), 1);
+  assert.equal(summary.reduce((n, c) => n + c.estimate, 0), 2);
   assert.equal(summary.reduce((n, c) => n + c.platform, 0), 16);
-  assert.equal(summary.reduce((n, c) => n + c.unknown, 0), 982);
+  assert.equal(summary.reduce((n, c) => n + c.unknown, 0), 983);
   assert.equal(snapshot.companies.find((c) => c.name === "Rork")?.revenueStatus, "unknown");
   assert.equal(snapshot.companies.find((c) => c.name === "Result")?.revenueStatus, "unknown");
   assert.ok(snapshot.observations.filter((o) => o.selection === "signal").every((o) => o.scope === "公司整体"));
@@ -97,7 +97,7 @@ test("disclosure counts exclude client income, biographies, targets and duplicat
 
 test("technical reviews keep accessible materials separate from tested capabilities", () => {
   const snapshot = structuredClone(seed) as ResearchSnapshot;
-  assert.equal(snapshot.companies.filter((c) => c.technicalReview).length, 22);
+  assert.equal(snapshot.companies.filter((c) => c.technicalReview).length, 27);
   assert.equal(snapshot.companies.filter((c) => c.technicalReview?.level === "public-artifact").length, 7);
   assert.equal(snapshot.companies.filter((c) => c.technicalReview?.level === "independent-test").length, 1);
   assert.equal(snapshot.companies.find((c) => c.id === "AI051")?.revenueStatus, "unknown");
@@ -307,7 +307,7 @@ test("unreconciled directory financials remain source reviews instead of income
   assert.ok(review.evidenceIds.includes("PH-TANA-DR"));
   assert.equal(reviewCompanyIds(snapshot, review).length, 0);
   assert.ok(!snapshot.observations.some((o) => o.evidence === "PH-TANA-DR"));
-  assert.equal(snapshot.candidateBatches!.flatMap((b) => b.records).filter((r) => r.status === "pending-year").length, 8);
+  assert.equal(snapshot.candidateBatches!.flatMap((b) => b.records).filter((r) => r.status === "pending-year").length, 9);
 });
 
 test("product pivots and founder histories do not become new company years or current revenue", () => {

content/research/ai-companies.seed.json

@@ -1,5 +1,5 @@
 {
-  "version": "2026-10-30-income-boundaries-2",
+  "version": "2026-10-30-cn-data-quality-1",
   "asOf": "2026-10-30",
   "scope": "全球公开资料研究：保留首批个案，并扩展核查2025—2026八个YC批次候选池；成立年份以官方公司页面字段为依据。",
   "limitations": [
@@ -2401,24 +2401,38 @@
       "business": "帮助企业构建强化学习环境和后训练数据集。",
       "buyer": "数据创业者、模型团队和AI实验室",
       "delivery": "公司自述有50多家企业采用；采用数不等于付费客户数或收入",
-      "notes": "目录主营与买家已人工复核；收入、定价与技术能力仍需专项核验。",
+      "notes": "官网卖家收益展示、任务数量和采用企业数均无公司收入凭证，不纳为HUD收入；平台佣金、环境销售和服务合同边界待核。",
       "cohort": "2025年成立",
       "teamBasis": "YC官方页面Team Size；2026-10-30访问快照，业务截至日未注明",
-      "price": "本轮未逐一核验",
-      "stage": "目录公开，商业成熟度待核验",
-      "sourceType": "YC官方页面中的公司自述",
+      "price": "未取得公开统一费率或合同；首页收益展示不是产品价格",
+      "stage": "产品平台及销售入口公开；实际付费规模和收入未知",
+      "sourceType": "YC及公司官网产品与收益展示补核",
       "evidenceIds": [
-        "YC-E-30254"
+        "YC-E-30254",
+        "HUD-HOME",
+        "HUD-PROFILE"
       ],
       "revenueStatus": "unknown",
-      "profileOnly": true,
+      "profileOnly": false,
       "classificationBasis": "2026-10-30人工复核YC官方业务介绍：帮助企业构建强化学习环境和后训练数据集。 按主要买家和工作任务归类；简介不足的推定已注明。",
       "directoryTags": [
         "Marketplace",
         "Reinforcement Learning",
         "B2B",
         "AI"
-      ]
+      ],
+      "directoryOrigin": "yc-2025-2026",
+      "businessModel": "RL环境开发/运行平台与数据销售入口；实际平台费、佣金及卖家/平台总净收入需按合同核实",
+      "entityBasis": "YC关联HUD品牌和hud.ai；DataVendor为官网销售入口，不据入口或卖家展示推独立法律主体或合并收入。",
+      "technicalReview": {
+        "level": "description",
+        "reviewedAt": "2026-10-30",
+        "summary": "已读环境构建、任务追踪和QA功能介绍；没有运行SDK、验证统计或测试奖励审计效果。",
+        "evidenceIds": [
+          "HUD-HOME",
+          "HUD-PROFILE"
+        ]
+      }
     },
     {
       "id": "YC30260",
@@ -4911,26 +4925,41 @@
       "year": 2025,
       "team": 8,
       "location": "San Francisco",
-      "business": "公司简介：Voice data for AI",
-      "buyer": "本轮未逐一核验",
-      "delivery": "公开YC公司资料；交付与付费使用待核验",
-      "notes": "目录基础资料；AI相关描述与业务分类均待进一步人工复核。收入、定价和技术能力未逐一核验。",
+      "business": "收集、标注并授权语音交互数据，为语音模型和Agent训练/评估提供数据。",
+      "buyer": "语音模型、语音Agent及多模态研发团队",
+      "delivery": "官网及YC说明语音数据和研究产品，实际合同、交付与估算收入未核。",
+      "notes": "只有第三方2025年年收入估算，不是公司自报ARR或当前收入；估算方法、币种、总额/净额及时间覆盖留缺。与同名服装企业分开。",
       "cohort": "2025年成立",
       "teamBasis": "YC官方页面Team Size；2026-10-30访问快照，业务截至日未注明",
-      "price": "本轮未逐一核验",
-      "stage": "目录公开，商业成熟度待核验",
-      "sourceType": "YC官方页面中的公司自述",
+      "price": "YC称灵活授权，官网提供数据采购联系；未取得统一价表",
+      "stage": "公开数据产品；收入仅第三方历史估算，未审计",
+      "sourceType": "YC与官方产品补核，第三方估算单列",
       "evidenceIds": [
-        "YC-E-30500"
+        "YC-E-30500",
+        "BS-PROFILE",
+        "BS-HOME",
+        "BS-LATKA"
       ],
-      "revenueStatus": "unknown",
-      "profileOnly": true,
-      "classificationBasis": "目录标签的规则初分",
+      "revenueStatus": "estimate",
+      "profileOnly": false,
+      "classificationBasis": "2026-10-30人工复核YC和官网：交付为语音数据与训练环境，保留C02；非直接语音模型收入。",
       "directoryTags": [
         "AIOps",
         "Data Labeling",
         "AI"
-      ]
+      ],
+      "directoryOrigin": "yc-2025-2026",
+      "businessModel": "语音数据授权、标注与训练环境；许可、项目服务与劳务成本需拆分，实际费率和经常性未知",
+      "entityBasis": "YC链接besimple.ai，业务与创始人对应；精确登记主体未核，不将同名服装企业财务并入。",
+      "technicalReview": {
+        "level": "description",
+        "reviewedAt": "2026-10-30",
+        "summary": "已读语音数据和研究自述；未检查数据集许可、下载材料或复现基准，不授予独立能力评级。",
+        "evidenceIds": [
+          "BS-HOME",
+          "BS-PROFILE"
+        ]
+      }
     },
     {
       "id": "YC30501",
@@ -5762,26 +5791,41 @@
       "year": 2025,
       "team": 4,
       "location": "San Francisco",
-      "business": "公司简介：Real-world company data for frontier AI",
-      "buyer": "本轮未逐一核验",
-      "delivery": "公开YC公司资料；交付与付费使用待核验",
-      "notes": "目录基础资料；AI相关描述与业务分类均待进一步人工复核。收入、定价和技术能力未逐一核验。",
+      "business": "帮助企业将运营数据及真实工作记录转为AI训练/评估数据并授权。",
+      "buyer": "AI实验室与模型团队；数据提供企业作为供给方",
+      "delivery": "官网介绍范围协商、数据处理和授权流程；实际合同、交付和收入未核。",
+      "notes": "客户潜在数据价值、提供方许可收入和Sureform平台收入不同；不把估价范围当公司收入，不与同名表单软件并表。",
       "cohort": "2025年成立",
       "teamBasis": "YC官方页面Team Size；2026-10-30访问快照，业务截至日未注明",
-      "price": "本轮未逐一核验",
-      "stage": "目录公开，商业成熟度待核验",
-      "sourceType": "YC官方页面中的公司自述",
+      "price": "估价取决于数据范围和需求，本轮未取得平台费率；客户估价不作平台订阅价",
+      "stage": "公开数据授权业务流程；具体成交与收入未知",
+      "sourceType": "YC及官方产品/条款补核",
       "evidenceIds": [
-        "YC-E-30561"
+        "YC-E-30561",
+        "SF-HOME",
+        "SF-PROFILE",
+        "SF-TERMS"
       ],
       "revenueStatus": "unknown",
-      "profileOnly": true,
-      "classificationBasis": "目录标签的规则初分",
+      "profileOnly": false,
+      "classificationBasis": "2026-10-30人工复核YC和官网：提供训练数据授权与处理，保留C02，不按机器人采集场景归本体厂商。",
       "directoryTags": [
         "Marketplace",
         "Data Labeling",
         "AI"
-      ]
+      ],
+      "directoryOrigin": "yc-2025-2026",
+      "businessModel": "连接企业数据供给与AI买方的数据授权业务；平台费、分成、数据成本和经常性需实际合同验证",
+      "entityBasis": "官网页脚列Atom Copilot, Corp.；YC链接Sureform品牌，法律沿革和成立时间未独立注册核验。",
+      "technicalReview": {
+        "level": "description",
+        "reviewedAt": "2026-10-30",
+        "summary": "已读采集、去标识和交付流程；未测试处理管道、读取数据集或验证安全承诺。",
+        "evidenceIds": [
+          "SF-HOME",
+          "SF-PROFILE"
+        ]
+      }
     },
     {
       "id": "YC30562",
@@ -29127,6 +29171,77 @@
           "ELORIAN-SUPPORT"
         ]
       }
+    },
+    {
+      "team": null,
+      "revenueStatus": "unknown",
+      "price": "本轮未取得统一公开价格或已签客户合同",
+      "sourceType": "中文发现线索与政府/公司公开原始资料补核",
+      "id": "AI083",
+      "name": "清析智研（CleerScience）",
+      "url": "https://www.cleerscience.com/",
+      "cat": "C07",
+      "year": 2026,
+      "location": "中国北京；政府调研所指企业",
+      "business": "科学智能与新材料研发，面向材料性质预测、表征分析和工艺优化。",
+      "buyer": "材料产业研发团队、高校和科研机构",
+      "delivery": "政府调研记录介绍产品与合作，自主实验室建设仍在推进；本轮未测试或核商业交付。",
+      "notes": "成立2026据北京市经信局调研中的企业介绍，精确注册日未知。官网读取留缺，36氪当前正文未确认检索中的公司名。融资、研发项目和合作不当收入。",
+      "cohort": "2026年成立（政府调研记录中的企业自述）；精确注册未核",
+      "teamBasis": "未取得完整员工数，不从创始人、合作团队或融资规模推人数",
+      "stage": "研发与产业合作公开；实际合同、收入及独立性能待核",
+      "evidenceIds": [
+        "KR-CLEER",
+        "CLEER-GOV",
+        "CLEER-HOME-ACCESS"
+      ],
+      "businessModel": "科学AI模型软件与研发解决方案；模型服务、实验及工程交付的合同与成本结构待核",
+      "entityBasis": "政府材料列北京清析智研科技有限公司；CleerScience品牌官网本轮不可读，网站主体归属与精确登记仍需补核。",
+      "classificationBasis": "按材料科研与工艺研发工作归C07；研发世界模型的愿景不自动等通用模型主营。",
+      "technicalReview": {
+        "level": "description",
+        "reviewedAt": "2026-10-30",
+        "summary": "已读政府调研中的企业技术介绍；没有官网完整材料、独立测试或可核代码/模型评审。",
+        "evidenceIds": [
+          "CLEER-GOV"
+        ]
+      }
+    },
+    {
+      "team": null,
+      "revenueStatus": "unknown",
+      "price": "本轮未取得统一公开价格或已签客户合同",
+      "sourceType": "中文发现线索与政府/公司公开原始资料补核",
+      "id": "AI084",
+      "name": "星源智",
+      "url": "https://xyz-eai.com/",
+      "cat": "C10",
+      "year": 2025,
+      "location": "中国北京亦庄；公开总部迁入信息",
+      "business": "为多类机器人提供具身大脑、世界模型及端侧算力/部署方案。",
+      "buyer": "机器人厂商及产业场景集成客户；具体付费客户未核",
+      "delivery": "官网披露具身大脑、算力产品及装卸方案发布；本轮未测试或核实际验收。",
+      "notes": "政府园区材料列2025年8月成立，官网自述8月1日；精确注册未独立查询。三年战略订单、融资和客户覆盖宣传不作当前收入或ARR。",
+      "cohort": "2025年8月成立；官网自述日为2025-08-01，注册未独立核验",
+      "teamBasis": "未取得完整员工数，不把技术团队背景或客户覆盖比例当团队人数",
+      "stage": "产品和场景方案公开；模型/硬件/集成的合同、验收和财务待核",
+      "evidenceIds": [
+        "KR-XYZ",
+        "XYZ-GOV",
+        "XYZ-ABOUT"
+      ],
+      "businessModel": "机器人具身大脑、端侧算力及方案交付；模型许可、硬件和集成服务需分业务核收入",
+      "entityBasis": "官网列北京星源智机器人科技有限公司，政府园区资料对应；不并入智源研究院、智谱或同名企业历史。",
+      "classificationBasis": "按机器人具身大脑与部署任务归C10，模型架构不是独立能力排名。",
+      "technicalReview": {
+        "level": "description",
+        "reviewedAt": "2026-10-30",
+        "summary": "已读具身大脑与端侧产品自述和公开园区介绍；没有独立模型/机器人测试。",
+        "evidenceIds": [
+          "XYZ-ABOUT",
+          "XYZ-GOV"
+        ]
+      }
     }
   ],
   "observations": [
@@ -32208,6 +32323,21 @@
       "reason": "原文未展开运行率周期、经常性及总额/净额定义；无明确指标截至日和币种，不补成年化收入或ARR，不进MRR比较组",
       "bound": ">",
       "comparisonGroup": null
+    },
+    {
+      "id": "BS-O-LATKA-2025",
+      "company": "YC30500",
+      "name": "Besimple AI",
+      "kind": "第三方估算年收入，非公司披露ARR",
+      "value": 440000,
+      "unit": "原文$，币种待核",
+      "timing": "估算标2025年9月；年度覆盖未展开，页面更新2026-08-10，2026-10-30读取",
+      "evidence": "BS-LATKA",
+      "scope": "公司整体",
+      "selection": "historical-estimate",
+      "reason": "第三方EST，未取得估算底表或原始收入凭据；非经常性ARR、非同期实报数，不进入可比样本",
+      "bound": "约",
+      "comparisonGroup": null
     }
   ],
   "evidence": [
@@ -47324,6 +47454,204 @@
       "kind": "公司官方条款局部读取",
       "summary": "直达页面可读部分列Allus AI Inc.、Atlanta及评估/试用范围。",
       "limitation": "直达部分正文呈模板占位，完整条款仅有检索返回，未标全文已核。不能据此确认第三方月费、实际付费试点或收入；责任上限不是营业收入。"
+    },
+    {
+      "id": "BS-PROFILE",
+      "companyId": "YC30500",
+      "title": "Besimple：公司与交付核查",
+      "url": "https://www.ycombinator.com/companies/besimple-ai",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "YC官方公司自述",
+      "summary": "YC关联besimple.ai、Yi Zhong及Bill Wang，描述语音数据收集、标注与授权。",
+      "limitation": "自报数据时长、创始人旧工作与用户效果不当收入；不同于同名印度服装或法国公司。"
+    },
+    {
+      "id": "BS-HOME",
+      "companyId": "YC30500",
+      "title": "Besimple：官方产品资料",
+      "url": "https://besimple.ai/",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "公司官方产品材料",
+      "summary": "官网提供语音交互数据、训练环境及研究介绍，并提供数据采购联系入口。",
+      "limitation": "未取得公司收入、统一公开价或客户合同；数据量、基准与模型训练声明未独立测试。"
+    },
+    {
+      "id": "BS-LATKA",
+      "companyId": "YC30500",
+      "title": "Besimple：第三方收入估算口径",
+      "url": "https://getlatka.com/companies/besimple.ai",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "第三方模型估算",
+      "summary": "公开页将2025年9月年收入估为约$440k，标EST并说明公司未公布自身收入；页面更新2026-08-10。",
+      "limitation": "不是公司披露ARR或审计数，未取得估算底表；币种和年度覆盖未展开。员工4人为较早目录时点，不覆盖当前YC团队8人；不把估算页与原始采访混用。"
+    },
+    {
+      "id": "HUD-HOME",
+      "companyId": "YC30254",
+      "title": "HUD：首页收益展示与公司收入区分",
+      "url": "https://www.hud.ai/",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "公司官方产品与展示资料",
+      "summary": "官网介绍环境SDK、任务追踪、QA及DataVendor销售入口，并展示卖家月度收益和销售金额。",
+      "limitation": "页面展示无可核卖家身份、交易凭证或公司整体收入证明；不将展示收益、任务运行量、推理调用量计入HUD收入或ARR。未运行SDK。"
+    },
+    {
+      "id": "HUD-PROFILE",
+      "companyId": "YC30254",
+      "title": "HUD：官方目录补查",
+      "url": "https://www.ycombinator.com/companies/hud",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "YC官方公司自述",
+      "summary": "YC描述RL环境/数据平台及50多家企业使用，关联hud.ai。",
+      "limitation": "采用企业数不等付费客户数，卖家收入与平台佣金需拆分，未取得当前合同、费率或实际收入。"
+    },
+    {
+      "id": "SF-HOME",
+      "companyId": "YC30561",
+      "title": "Sureform：客户数据价值不等平台收入",
+      "url": "https://www.sureformhq.com/",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "公司官方产品资料",
+      "summary": "官网协助企业按范围授权运营数据，给出客户数据可能价值$50k—$4m+，收入归属描述为数据提供企业获许可款。",
+      "limitation": "潜在数据估价不是Sureform收入或成交额；不是保证报价、合同或ARR，不按范围中点或上限估收入。"
+    },
+    {
+      "id": "SF-PROFILE",
+      "companyId": "YC30561",
+      "title": "Sureform：目录与产品沿革",
+      "url": "https://www.ycombinator.com/companies/sureform",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "YC官方公司自述",
+      "summary": "YC描述企业运营数据与机器人/世界模型训练数据业务，关联Sureform创始人及sureformhq.com。",
+      "limitation": "旧发布的实体数据采集与当前运营数据授权分别保留，未证实收入结构变化；不与sureform.app表单软件混用。"
+    },
+    {
+      "id": "SF-TERMS",
+      "companyId": "YC30561",
+      "title": "Sureform：平台条款与实体线索",
+      "url": "https://www.sureformhq.com/terms",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "公司官方合同资料",
+      "summary": "官方条款入口与页脚关联Atom Copilot, Corp.，网站用于洽谈数据授权合作。",
+      "limitation": "本轮未取得双方签署商业合同或费率，不从条款发布日期推法律成立日，Sureform品牌与实体沿革仍需登记补核。"
+    },
+    {
+      "id": "KR-CLEER",
+      "companyId": "AI083",
+      "title": "清析智研：36氪检索发现与正文差异",
+      "url": "https://www.36kr.com/p/4043538991797379",
+      "publishedAt": "2026-10-27",
+      "accessedAt": "2026-10-30",
+      "kind": "中文媒体检索发现",
+      "summary": "检索返回清析智研线索，但当前直达文章未找到公司名。",
+      "limitation": "不声称原文已确认该条目；成立与业务另据独立政府调研材料，媒体关联继续留缺。"
+    },
+    {
+      "id": "KR-XYZ",
+      "companyId": "AI084",
+      "title": "星源智：中文媒体发现",
+      "url": "https://eu.36kr.com/zh/p/3953252758748295",
+      "publishedAt": "2026-08-24",
+      "accessedAt": "2026-10-30",
+      "kind": "中文媒体公开授权转载",
+      "summary": "36氪欧洲站可读中国企业家杂志文章的星源智相关段落。",
+      "limitation": "转载与公司原始声明分开；融资、出货、POC、市场预测和效果宣传不是收入，未读遍两年档案。"
+    },
+    {
+      "id": "KR-SIMPLE",
+      "companyId": null,
+      "title": "深朴智能：报道与登记年份差异",
+      "url": "https://www.36kr.com/p/4043850978367360",
+      "publishedAt": "2026-10-28",
+      "accessedAt": "2026-10-30",
+      "kind": "中文媒体平台发表文章",
+      "summary": "芯智能头条在36氪发布的文章把深朴智能列为2025成立。",
+      "limitation": "以交易当事方原始公告所列2024登记另核，不将融资或产品发布当新成立；不复刻20家公司清单。"
+    },
+    {
+      "id": "KR-AIMOGA",
+      "companyId": null,
+      "title": "墨甲机器人：中文媒体发现",
+      "url": "https://eu.36kr.com/zh/p/3954273153586820",
+      "publishedAt": "2026-08-25",
+      "accessedAt": "2026-10-30",
+      "kind": "中文媒体公开转载",
+      "summary": "文章将墨甲公司成立与机器人交付、售价、产能并列介绍。",
+      "limitation": "仅作候选发现；公司、母公司与早期机器人项目沿革未补核，交付量乘价格不是收入或ARR。"
+    },
+    {
+      "id": "CLEER-GOV",
+      "companyId": "AI083",
+      "title": "清析智研：政府原始调研记录",
+      "url": "https://jxj.beijing.gov.cn/ztzl/ywzt/hbjh/hbdt/hbdt/202609/t20260918_4870458.html",
+      "publishedAt": "2026-09-18",
+      "accessedAt": "2026-10-30",
+      "kind": "政府官方调研记录",
+      "summary": "北京市经信局调研北京清析智研科技有限公司，记录企业自述2026成立及材料预测、表征和工艺优化业务。",
+      "limitation": "政府调研中的企业介绍不是工商独立审计或技术评测，精确成立日、商业合同、团队和收入未知。"
+    },
+    {
+      "id": "CLEER-HOME-ACCESS",
+      "companyId": "AI083",
+      "title": "清析智研：官网访问缺项",
+      "url": "https://www.cleerscience.com/",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "公司官方入口读取核查",
+      "summary": "官网本轮读取超时；品牌新闻页亦无可读正文。",
+      "limitation": "不记为已读官网或完整法律主体核验，当前成立与业务以政府调研记录为依据。"
+    },
+    {
+      "id": "XYZ-GOV",
+      "companyId": "AI084",
+      "title": "星源智：政府公开园区记录",
+      "url": "https://kw.beijing.gov.cn/xwdt/kcyx/xwdtyqqy/202610/t20261014_4900927.html",
+      "publishedAt": "2026-10-14",
+      "accessedAt": "2026-10-30",
+      "kind": "政府官方园区资料",
+      "summary": "北京市科委公开园区报道列北京星源智机器人科技有限公司2025年8月成立、机器人具身大脑及端侧部署业务。",
+      "limitation": "页面注明来源亦城时报；成立为公开介绍非工商底档，融资、覆盖比例和演示记录不作公司收入或独立能力等级。"
+    },
+    {
+      "id": "XYZ-ABOUT",
+      "companyId": "AI084",
+      "title": "星源智：公司官方沿革",
+      "url": "https://xyz-eai.com/about.html",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "公司官方成立与产品资料",
+      "summary": "官网自述2025-08-01成立，介绍通用具身大脑、端侧算力及装卸方案。",
+      "limitation": "精确日为公司自述未独立核登记；战略三年订单目标、融资和产品发布不等已确认收入或ARR。"
+    },
+    {
+      "id": "SIMPLE-CNINFO",
+      "companyId": null,
+      "title": "深朴智能：交易当事方工商资料",
+      "url": "https://static.cninfo.com.cn/finalpage/2026-09-23/1225577942.PDF",
+      "publishedAt": "2026-09-23",
+      "accessedAt": "2026-10-30",
+      "kind": "交易当事方原始公告PDF",
+      "summary": "美芝股份重整投资协议公告第4页列北京深朴智能科技有限公司及成立日期2024-10-18。",
+      "limitation": "只用相关页文字核登记线索，不声称读完41页或独立工商查询；标的上市公司财务、重组保证金和出资额不当深朴收入。"
+    },
+    {
+      "id": "AIMOGA-HOME",
+      "companyId": null,
+      "title": "墨甲机器人：公司官方产品与主体入口",
+      "url": "https://www.aimoga-robot.com/cn/zh/",
+      "publishedAt": null,
+      "accessedAt": "2026-10-30",
+      "kind": "公司官方产品资料",
+      "summary": "官网展示人形/四足产品，页脚列安徽墨甲智创机器人科技有限公司。",
+      "limitation": "本轮可读入口未取得原始成立日期；母公司历史、出货、售价、产能和计划不直接填入收入。"
     }
   ],
   "reviews": [
@@ -48990,6 +49318,99 @@
         "ALLUS-TERMS"
       ],
       "handling": "仅有第三方模糊宣传和局部官方条款；不补数值，不用工厂月费反推客户，保持收入未知。 该核查不单独证明财务造假，也不是投资或审计结论。"
+    },
+    {
+      "id": "BS-R-ESTIMATE",
+      "type": "数据供应商收入与归属口径核查",
+      "company": "Besimple AI",
+      "scope": "companies",
+      "companyIds": [
+        "YC30500"
+      ],
+      "summary": "语音数据授权、标注与训练环境；许可、项目服务与劳务成本需拆分，实际费率和经常性未知",
+      "evidenceIds": [
+        "BS-PROFILE",
+        "BS-HOME",
+        "BS-LATKA"
+      ],
+      "handling": "估计值保留为历史第三方年收入，时间、币种和方法留缺；不当ARR或十月现状，不转移其他同名公司财务。 该核查不单独证明财务造假。"
+    },
+    {
+      "id": "HUD-R-EARNINGS",
+      "type": "数据供应商收入与归属口径核查",
+      "company": "HUD",
+      "scope": "companies",
+      "companyIds": [
+        "YC30254"
+      ],
+      "summary": "RL环境开发/运行平台与数据销售入口；实际平台费、佣金及卖家/平台总净收入需按合同核实",
+      "evidenceIds": [
+        "HUD-HOME",
+        "HUD-PROFILE"
+      ],
+      "handling": "卖家收益展示、任务与调用量不当平台收入；实际费率、真实交易和平台/卖家总净口径待核，未新增收入观测。 该核查不单独证明财务造假。"
+    },
+    {
+      "id": "SF-R-VALUATION",
+      "type": "数据供应商收入与归属口径核查",
+      "company": "Sureform",
+      "scope": "companies",
+      "companyIds": [
+        "YC30561"
+      ],
+      "summary": "连接企业数据供给与AI买方的数据授权业务；平台费、分成、数据成本和经常性需实际合同验证",
+      "evidenceIds": [
+        "SF-HOME",
+        "SF-PROFILE",
+        "SF-TERMS"
+      ],
+      "handling": "数据供给方估价与平台收入分开；不能从潜在价值范围反推收入、毛利或ARR，实际合同待核。 该核查不单独证明财务造假。"
+    },
+    {
+      "id": "AI083-R-SCOPE",
+      "type": "中文新公司成立与商业边界核查",
+      "company": "清析智研（CleerScience）",
+      "scope": "companies",
+      "companyIds": [
+        "AI083"
+      ],
+      "summary": "科学AI模型软件与研发解决方案；模型服务、实验及工程交付的合同与成本结构待核",
+      "evidenceIds": [
+        "KR-CLEER",
+        "CLEER-GOV",
+        "CLEER-HOME-ACCESS"
+      ],
+      "handling": "成立2026据北京市经信局调研中的企业介绍，精确注册日未知。官网读取留缺，36氪当前正文未确认检索中的公司名。融资、研发项目和合作不当收入。 未取得可用公司收入，不新增数值观测。"
+    },
+    {
+      "id": "AI084-R-SCOPE",
+      "type": "中文新公司成立与商业边界核查",
+      "company": "星源智",
+      "scope": "companies",
+      "companyIds": [
+        "AI084"
+      ],
+      "summary": "机器人具身大脑、端侧算力及方案交付；模型许可、硬件和集成服务需分业务核收入",
+      "evidenceIds": [
+        "KR-XYZ",
+        "XYZ-GOV",
+        "XYZ-ABOUT"
+      ],
+      "handling": "政府园区材料列2025年8月成立，官网自述8月1日；精确注册未独立查询。三年战略订单、融资和客户覆盖宣传不作当前收入或ARR。 未取得可用公司收入，不新增数值观测。"
+    },
+    {
+      "id": "KR-R-CANDIDATE-GAPS",
+      "type": "中文媒体年份与主体缺项复核",
+      "company": "深朴智能及墨甲机器人候选",
+      "scope": "excluded-candidate",
+      "summary": "深朴据交易当事方登记资料归2024既有公司，墨甲原始成立和母公司项目沿革继续待核。",
+      "evidenceIds": [
+        "KR-SIMPLE",
+        "SIMPLE-CNINFO",
+        "KR-AIMOGA",
+        "AIMOGA-HOME"
+      ],
+      "handling": "媒体与工商日期分别保存，不将融资、重组、品牌或机器人发布当新成立。发现页可读不代表所有候选完成财务或技术审查。"
     }
   ],
   "coverage": {
@@ -50325,6 +50746,76 @@
           "note": "MIT托管创始人履历第1页列2024年5月联合创办Eka Robotics，按早于2025的既有业务排除；不是精确注册证明。此前访问失败记录保留，不把2026融资或模型发布当新成立。"
         }
       ]
+    },
+    {
+      "id": "KR-2026-targeted-1",
+      "sourceId": "S11",
+      "url": "https://www.36kr.com/p/4043538991797379",
+      "accessedAt": "2026-10-30",
+      "scope": "四项定向检索与相关段落补查：2项据独立原始资料关联公司，1项范围外，1项成立待核；清析智研与发现文章的正文关联未确认，保留检索/直达差异。不是四篇全文、榜单或两年档案覆盖。",
+      "complete": false,
+      "records": [
+        {
+          "id": "KR26-01",
+          "name": "清析智研",
+          "url": "https://www.36kr.com/p/4043538991797379",
+          "categoryId": "C07",
+          "kind": "company",
+          "profileStatus": "archive-only",
+          "status": "linked",
+          "companyId": "AI083",
+          "evidenceIds": [
+            "KR-CLEER",
+            "CLEER-GOV",
+            "CLEER-HOME-ACCESS"
+          ],
+          "note": "检索摘要发现但当前直达未找到公司名；成立及业务由独立政府调研补证，官网不可读和媒体关联留缺。"
+        },
+        {
+          "id": "KR26-02",
+          "name": "星源智",
+          "url": "https://eu.36kr.com/zh/p/3953252758748295",
+          "categoryId": "C10",
+          "kind": "company",
+          "profileStatus": "readable",
+          "status": "linked",
+          "companyId": "AI084",
+          "evidenceIds": [
+            "KR-XYZ",
+            "XYZ-GOV",
+            "XYZ-ABOUT"
+          ],
+          "note": "中文转载相关段落可读，成立另据政府与公司原始沿革；融资、三年目标和发布不当收入。"
+        },
+        {
+          "id": "KR26-03",
+          "name": "深朴智能",
+          "url": "https://www.36kr.com/p/4043850978367360",
+          "categoryId": "C10",
+          "kind": "company",
+          "profileStatus": "readable",
+          "status": "outside-cohort",
+          "evidenceIds": [
+            "KR-SIMPLE",
+            "SIMPLE-CNINFO"
+          ],
+          "note": "文章列2025，交易当事方原始公告列2024-10-18登记，按范围外公司排除；保留年份差异，不把上市标的财务归新业务。"
+        },
+        {
+          "id": "KR26-04",
+          "name": "墨甲机器人（AiMOGA）",
+          "url": "https://eu.36kr.com/zh/p/3954273153586820",
+          "categoryId": "C10",
+          "kind": "company",
+          "profileStatus": "readable",
+          "status": "pending-year",
+          "evidenceIds": [
+            "KR-AIMOGA",
+            "AIMOGA-HOME"
+          ],
+          "note": "官网产品与运营名称可读，媒体2025成立尚待原始登记/沿革补核；母公司项目史、出货乘售价、产能及目标不作收入。"
+        }
+      ]
     }
   ],
   "directoryRechecks": [

content/research/ai-companies.sources.json

@@ -1,5 +1,5 @@
 {
-  "version": "2026-10-30-26",
+  "version": "2026-10-30-27",
   "scopeConfirmedAt": "2026-10-30",
   "acceptance": "以确认的公开来源清单为边界，逐条核对并标明缺项。",
   "provenance": "S01—S12恢复自已提交用户的《AI公司研究_来源与分类模板》来源清单；S13—S16单列后续已使用的补充证据渠道，不将使用次数当作全量覆盖。",
@@ -15,7 +15,7 @@
       ],
       "initialAccess": "官网目录可访问",
       "boundary": "已读取八个2025—2026批次镜像候选池1428条，再核YC官方页面；不声称覆盖全部YC历年或未公开公司。",
-      "nextCheck": "本轮3条原目录记录深化为商业/收入个案并保留原来源；当前目录基础项176条主营分类人工复核。继续核其余规则初分及财务缺项，不把分类、收入运行率或月度合同直接当可比ARR。",
+      "nextCheck": "原新增目录中22条深化为个案并保留来源；当前基础项175条主营分类人工复核。继续核其余资料与财务缺项，估算、卖家收益展示和数据供给方估价不当公司ARR。",
       "limitation": "批次不等于成立年；不代表全部公司"
     },
     {
@@ -214,15 +214,16 @@
       "name": "36氪",
       "url": "https://www.36kr.com/",
       "role": "reporting",
-      "status": "access-limited",
+      "status": "targeted",
       "hosts": [
         "www.36kr.com",
-        "36kr.com"
+        "36kr.com",
+        "eu.36kr.com"
       ],
       "initialAccess": "安全检测，内容未核验",
-      "boundary": "2026-10-30仍返回安全检测页面，未把首页作为已读文章。",
-      "nextCheck": "使用可公开读取的具体文章或其他原始来源；不绕过访问保护。",
-      "limitation": "暂不作为已读取内容证据",
+      "boundary": "此前首页安全检测仍保留；本轮4项定向检索及相关段落补查，2项据政府/公司材料关联、1项按原始登记归范围外、1项年份待核。清析检索与直达正文不一致未确认关联，其他文章为平台发布/授权转载，不代表两年全量。",
+      "nextCheck": "补核墨甲原始成立与母公司项目沿革、清析发现文章关联及未读正文；继续建两年公开文章台账，不绕过访问保护。",
+      "limitation": "仅核具体相关段落，检索和直达差异、转载与原始披露分别记录。",
       "accessRecheckedAt": "2026-10-30"
     },
     {
@@ -285,9 +286,11 @@
       "url": null,
       "role": "supplement",
       "status": "targeted",
-      "hosts": [],
+      "hosts": [
+        "getlatka.com"
+      ],
       "initialAccess": "见具体证据访问记录",
-      "boundary": "未归入上述来源的公开材料按原始证据链接逐条保留；不视为新增已批准的完整候选池。",
+      "boundary": "未归入上述来源的公开材料按原始证据链接逐条保留；不视为新增已批准的完整候选池。 本轮读取GetLatka单页历史年收入估算，方法底表未取得，不视为新增完整候选池。",
       "nextCheck": "逐项核来源归属、可靠性及可读正文，不能将搜索摘要当全文。",
       "limitation": "来源可读不等于声明已独立验证。"
     }

content/research/ai-income-analysis.json

@@ -1,5 +1,5 @@
 {
-  "sourceVersion": "2026-10-30-income-boundaries-2",
+  "sourceVersion": "2026-10-30-cn-data-quality-1",
   "asOf": "2026-10-30",
   "metric": "platform-mrr-usd",
   "sampleUnit": "product-or-business",