FairLab Pro 是 Tian-2 旗下面向高水平中学科研竞赛的 AI 辅导产品家族，覆盖 ISEF、丘成桐中学科学奖、Regeneron STS 和 Conrad Challenge 四条产品线。它帮助教练和学生在投入数月之前发现高成本问题：题目与既有工作的重合、缺失的审批、薄弱的对照、数据泄漏、论断缺乏支持，以及无法验证的学生所有权。FairLab Pro is Tian-2's family of AI coaching products for high-level student research competitions, with editions for ISEF, the S.-T. Yau High School Science Award, Regeneron STS, and the Conrad Challenge. It helps coaches and students catch expensive problems before months are invested: topic overlap with prior work, missing approvals, weak controls, data leakage, unsupported claims, and student ownership that cannot be verified.
每个分支都有一个 MIT 许可的免费公开 Research Suite 承担教育与入门；Pro 分支承载该比赛独有的规则、评测、时间线与工作流。分支之间不共享未经验证的规则和数据许可。Every edition ships a free, MIT-licensed public Research Suite for education and onboarding; the Pro edition carries that competition's own rules, evaluations, timelines, and workflows. Editions never share unverified rules or data licenses.
面向 ISEF 及其关联赛：SRC/IRB 分诊、Forms 与时序证据、题目情报、可复现性审计、答辩模拟。For ISEF and affiliated fairs: SRC/IRB triage, forms and timeline evidence, topic intelligence, reproducibility audits, viva practice.
面向丘奖：学科赛道导航、研究论文、导师/学生贡献边界、AI 使用合规、全英文答辩。For the S.-T. Yau High School Science Award: subject-track navigation, research papers, mentor/student contribution boundaries, AI-use compliance, English defense.
面向 STS：申请材料、研究报告、贡献归属、推荐信协调与评审路径。For Regeneron STS: application materials, research reports, contribution attribution, recommendation coordination, and the review pathway.
面向 Conrad：Innovation Brief、Lean Canvas、市场与财务、IP、视频与 Summit pitch。For the Conrad Challenge: Innovation Brief, Lean Canvas, market and finance, IP, video, and the Summit pitch.
付费价值来自持续维护的规则、已许可或用户提供的数据、评测、项目状态、组织工作流和人工复核——不来自隐藏提示词。复制提示词不应能够复现完整的付费产品。Paid value comes from continuously maintained rules, licensed or user-provided data, evaluations, project state, organizational workflows, and human review — not from hidden prompts. Copying prompts should never reproduce the full paid product.
以 ISEFLab Pro 当前实现为例。每项能力都有对应的评测用例守边界；报告注明规则、索引、账本和评测版本。As currently implemented in ISEFLab Pro. Each capability is guarded by evaluation cases; every report states its rules, index, ledger, and evaluation versions.
navigator告诉你当前项目最该先解决哪一道门禁：审批、设计、所有权还是提交物。Tells you which gate to clear next: approval, design, ownership, or submission.
topic-intelligence在权利许可或用户自带（BYO/本地）模式下检查题目与既有工作的重合，来源和权利模式可追溯。Checks topic overlap against prior work in rights-licensed or bring-your-own local modes, with traceable sources and rights modes.
compliance-ledger按当前赛季分诊规则、表单、审批、成人支持和 AI 使用披露，逐项落进版本化账本。Triage of current-cycle rules, forms, approvals, adult support, and AI-use disclosure, each recorded in a versioned ledger.
design-auditor审计科学与工程测试设计：变量隔离、对照、样本量与可行性。Audits scientific and engineering test design: variable isolation, controls, sample size, feasibility.
reproducibility-auditor审阅数据集、代码与统计描述：数据泄漏、受试者重复、论断是否被证据支持。Reviews datasets, code, and statistical descriptions: leakage, subject duplication, and whether claims are supported.
submission-critic只对学生已写出的实质草稿做批评和差异建议——从不生成初始提交材料。Critiques a substantive student draft and proposes diffs — never generates initial submission materials.
viva-simulator一次一问的评委式追问，在学生作答之前不展示模型答案。One-question-at-a-time judging practice that hides model answers until the student has responded.
ledger + gates展示每个项目的当前门禁、阻塞项、严重程度、负责人与下一行动，中英文输出。Shows each project's current gate, blockers, severity, owner, and next action, in Chinese and English.
定价尚未确定：在记录真实交付成本和用户反馈之前，不发布固定价格。当前以 ISEFLab Pro 为试点，由 Tian 人工复核首轮交付。Pricing is not yet set: no fixed prices will be published before real delivery costs and user feedback are on record. The current pilot runs on ISEFLab Pro with first-round deliveries reviewed by a human.