陪跑一个项目、一个赛季：三件套审计在赛季内可重复运行，整改表在原表上更新状态，把审批时序、证据缺口与研究质量问题一轮轮清掉，直到关联赛提交截止。One project, one season: the three-report audit can be re-run throughout the season and the remediation table is updated in place, clearing approval-timing problems, evidence gaps, and research-quality defects round by round until the affiliated-fair deadline.
¥7,999赛季 · 首轮试点免费per season · first pilot round free
为谁而设Who it's for
项目还在做，审批与证据的时序还来得及补The project is still running — the approval timing can still be fixed
ISEF 的多数硬性要求不是"截止日前交齐材料"，而是事前：人类被试、脊椎动物、潜在危险生物剂或组织的研究，须在招募、干预或采集数据开始之前拿到责任 IRB / SRC / IACUC / IBC 的裁定；Form 1、Form 1A 与原始 Research Plan 须在实验之前完成。这类时序一旦走反，能否补救由责任机构裁定。越早发现，可选项越多。Most of ISEF's hard requirements are not "submit everything before the deadline" — they are prior: studies involving human participants, vertebrate animals, potentially hazardous biological agents, or tissue need a determination from the responsible IRB / SRC / IACUC / IBC before recruitment, intervention, or data collection begins, and Form 1, Form 1A, and the original Research Plan must be complete before experimentation. Once that ordering is broken, whether it can be remedied is the responsible body's call — and the earlier it surfaces, the more options remain.
学生赛季版做的就是这件事：照 2027 官方规则把项目的材料和内容逐项过一遍（16 条硬性要求，每条注明官方出处和适用时点），把每个缺口挂上可检索的规则编号，并合并成一张"谁、在哪天之前、补什么、怎么算补完"的整改行动表；赛季内反复跑，表上的状态列一路更新到提交。That is what the Student tier does: it walks the project's materials and content against the 2027 official rules (16 hard requirements, each with its official source and applicable timing), attaches a resolvable rule id to every gap, and merges them into one table of who fixes what, by when, and how "fixed" is verified — re-run across the season, with the status column carried all the way to submission.
适合独立做项目的学生与家庭、第一次走关联赛流程的参赛者，以及想给学生一份外部独立把关的家长。名下有多个项目要一起管的教练、工作室或学校，请看 机构学校版。A good fit for students and families running a project independently, first-time entrants navigating the affiliated-fair process, and parents who want one independent external check. Coaches, studios, or schools managing several projects at once should see the Institution tier.
该不该选Fit check
什么情况该选它，什么情况不该选When to choose it — and when not to
该选的情况：学生已经写出实质初稿（不必完美，能看出研究问题、方法和证据安排就行）；关联赛归属和 Adult Sponsor 已经明确；愿意在赛季里按整改清单逐项补齐、改完回来复验。Choose this tier when a substantive student draft already exists (it needn’t be polished — the research question, method and evidence plan just have to be visible), the affiliated fair and Adult Sponsor are settled, and you’re willing to work down the remediation list during the season and come back for re-checks.
学生还没有实质初稿：先别选。内容层审计不会启动，只会返回问题清单与写作练习；这个阶段最要紧的，是把研究计划自己写出来。No substantive draft yet? Don’t choose it yet. The content audit won’t start; it returns a question list and writing exercises instead. At that stage nothing matters more than the student writing the research plan themselves.
需要多项目名册、批量队列或给负责人的汇总视图：不在本档范围，请看 机构学校版。Need a multi-project roster, batch queue or a coordinator’s summary view? That’s outside this tier — see Institution.
交付物Deliverables
三件套 + 按需菜单 O1–O8，三种格式The three-report set plus on-request menu O1–O8, in three formats
每次交付都是三份一起给：M1、M2、M3，HTML / PDF / DOCX 三种格式都有；每条结论都注明对应的规则编号和官方出处。Every delivery ships all three — M1, M2, and M3 — in HTML, PDF, and DOCX, with every finding carrying its rule id and official source.
M1 · 内容层审计报告M1 · Content-layer audit report
Research Plan 按官方要素逐项对照：rationale、研究问题与假设（工程项目为工程目标与预期结果）、materials、procedures、risk and safety、data analysis、bibliographyThe Research Plan checked element by element: rationale, research question and hypothesis (or engineering goal and expected outcomes), materials, procedures, risk and safety, data analysis, bibliography
主张—证据匹配：每条结论 → 对应结果 → 对应原始数据与分析代码；断链、分析单位错配、泄漏、伪重复、事后分组逐条定位到章节 / 表号 / 图号 / 代码行Claim-to-evidence matching: each conclusion → its result → its raw data and analysis code; broken links, analysis-unit mismatches, leakage, pseudoreplication, and post-hoc grouping each pinned to a section / table / figure / code line
实验与工程设计审计（变量与对照、测量效度、复现次数、失败判据），每条缺陷附可观察的完成判据Experimental and engineering design audit (variables and controls, measurement validity, replication, failure criteria), each defect paired with an observable completion criterion
学生所有权风险点 5–8 条：答辩最可能被追问且当前证据最薄弱的段落；做得好的部分如实写下来供保留Five to eight student-ownership risk points: the passages most likely to be probed with the weakest current evidence; strengths recorded honestly so they survive revision
学生没有实质初稿时 M1 不启动，改为返回问题清单、要件清单与写作练习With no substantive student draft, M1 does not start — it is replaced by a question list, a requirements list, and writing exercises
M2 · 材料层缺项报告M2 · Materials-layer gap report
2027 表单矩阵与事前审批时序逐项判定，每项附规则编号与该规则的适用时序：五件套时序、Research Plan 自撰与时序、AI 使用与 Form 2A 披露、事前审批触发、非豁免人类被试的 IRB、禁止的临床动作、自测例外、当年表单编号、逐表单时序、延续项目、各校外场所的 Form 2C、变更再审批、参赛资格、团队状态、关联赛后方法学冻结、Rules Wizard 的定位Item-by-item verdicts on the 2027 form matrix and prior-approval timing, each carrying its rule id and that rule's applicable timing: the five-piece form set, Research Plan authorship and timing, AI use and Form 2A disclosure, prior-approval triggers, IRB review for non-exempt human-participant studies, prohibited clinical actions, the self-testing exception, current-cycle form numbering, per-form timing, continuation projects, a Form 2C per external site, re-approval after changes, eligibility, team status, post-fair methodology freeze, and the place of the Rules Wizard
缺口排序：关键合规与事前审批 → 高严重度 → 内容与分析缺陷 → 形式问题Gaps ordered: critical compliance and prior approvals → high severity → content and analysis defects → formal issues
关联赛特有的摘要字数、版式与提交窗口不在规则基线内，一律标 unavailable 并指向官方页面，由学生与 Adult Sponsor 核验Affiliated-fair-specific abstract limits, layout, and submission windows are outside the rules baseline and are marked unavailable, pointing to the official page for the student and Adult Sponsor to verify
M3 · 整改行动与验收清单M3 · Remediation actions and acceptance checklist
M1、M2 全部 must-fix 合并成一张六列执行表：问题（含证据定位与规则编号）/ 应补内容 / 负责人（角色，不写姓名）/ 最晚日期 / 验收标准 / 状态Every must-fix from M1 and M2 merged into one six-column table: issue (with evidence location and rule id) / what to add / owner (a role, never a name) / latest date / acceptance criterion / status
以该项目最近生效的关联赛截止为 D-0 的倒排；晋级 ISEF 后以 ISEF 截止为第二锚点A countdown anchored to the project's next effective affiliated-fair deadline as D-0; once advanced to ISEF, the ISEF deadline becomes the second anchor
复审在原表上更新状态列，不另起新表；历史轮次留档不删除，这张表同时是验收记录Re-audits update the status column in place rather than starting a new table; historical rounds are archived, never deleted — the table doubles as the acceptance record
按需菜单 O1–O8（本档全菜单，单项目）On-request menu O1–O8 (full menu at this tier, single project)
O1 研究计划整改工作表 O2 表单与事前审批时序倒排表 O3 Form 2A 支持与 AI 使用事实采集表O1 Research-plan remediation worksheet O2 Form and prior-approval timing countdown O3 Form 2A support and AI-use fact sheet
O4 摘要与主张—证据核对表 O5 数据谱系与泄漏自检表 O6 题目重合与差异化证据表O4 Abstract and claim-to-evidence checklist O5 Data-provenance and leakage self-check O6 Topic-overlap and differentiation evidence table
O7 关联赛资格与截止日核验单 O8 答辩所有权问答清单与复盘O7 Affiliated-fair eligibility and deadline verification sheet O8 Viva ownership question list and debrief
O 项全部由学生或责任成人填写事实，我们只给结构与核验点，绝不代填Every O item is filled in with facts by the student or a responsible adult; we supply the structure and the verification points and never auto-fill
八关证据台账 + 答辩演练Eight-gate evidence ledger + viva practice
从收材料到答辩，八个关口的进度都记在证据台账上；台账只记状态和材料线索，不存学生的敏感信息Eight gates from intake to interview are tracked in the evidence ledger; it records only status and evidence pointers, never sensitive details
自适应评委答辩模拟：一次一问、所有权追问、观察式复盘；不给标准答案、不打分、不评估晋级前景Adaptive judge simulation: one question at a time, ownership follow-ups, observational debrief; no model answers, no scoring, no assessment of advancement prospects
流程Process
四步，从来信到交付Four steps from first email to delivery
来信Write in邮件说明身份、项目类别与所处阶段、打算参加哪一个关联赛及该赛截止，注明申请 学生赛季版。Email us who you are, the project's category and stage, which affiliated fair you plan to enter and its deadline, requesting the Student tier.
资格确认Eligibility confirmation核对三件事：项目有 Adult Sponsor、关联赛归属明确、初稿由学生本人撰写；名额和这一单怎么做，都在回信里确认。We verify three facts: the project has an Adult Sponsor, the affiliated fair is identified, and the student writes the initial drafts; availability and the terms of your order are confirmed by email.
材料收取Materials intake收学生初稿与表单 / 审批的状态（不是原件）。签名扫描件、知情同意书正文与伦理审查原件只在本地临时核对，不留副本。We collect the student's drafts and the status of forms and approvals — not the originals. Signature scans, consent text, and original ethics documents are checked locally and transiently, with no copies retained.
交付与复审Delivery and re-audit交付 M1 / M2 / M3 与授权 O 项，HTML / PDF / DOCX 三格式；赛季内按进度重跑，M3 在原表上更新状态。You receive M1 / M2 / M3 plus authorized O items in HTML / PDF / DOCX; audits are re-run as the season progresses, with M3 updated in place.
不包含What's not included
诚实清单：这一档没有的东西The honest list: what this tier does not have
多项目名册 / 批量交付队列 / 协调人与政策视图：仅 机构学校版Multi-project roster / batch delivery queue / coordinator and policy views — Institution tier only
人工教育性复核默认不含，可选加购；未加购时交付摘要如实注明"未含人工复核"The human educational review is not included by default — it is an optional add-on, and without it the delivery summary states "no human review included"
往届项目的完整摘要库（Society 受限数据）：我们没有转发的权利，任何档位、任何价格都不提供检索或转发；没有获许可的数据时，也不会输出重合计数或相似度Retrieval or redistribution of the Society restricted corpus (full abstracts and the project-level index) — we hold no redistribution rights, so this is unavailable in every tier at any price, and with no licensed corpus searched, no overlap count or similarity score is output
关联赛特有的摘要字数、展板版式与提交窗口：不在规则基线内，标 unavailable 由你与 Adult Sponsor 在官方页面核验Affiliated-fair-specific abstract limits, board layout, and submission windows — outside the rules baseline, marked unavailable for you and your Adult Sponsor to verify on the official page
任何官方角色：我们不担任 Adult Sponsor、Qualified Scientist、Designated Supervisor，也不代表你与 SRC / IRB 交涉Any official role: we do not act as Adult Sponsor, Qualified Scientist, or Designated Supervisor, and we do not deal with an SRC or IRB on your behalf
以及不论哪个档位、花多少钱都不提供的：代写研究计划 / 摘要 / 海报文本 / 论文初稿；晋级或获奖预测。这两条是红线，不是待解锁的高级功能。人工教育性复核是教育性复核，不替代 SRC、IRB、IACUC、IBC、Qualified Scientist、Adult Sponsor 或评委的任何裁定。And what no tier has at any price: ghostwritten research plans / abstracts / poster text / paper drafts, and advancement or award prediction. These are red lines, not locked premium features. The human educational review is an educational review and substitutes for no determination by an SRC, IRB, IACUC, IBC, Qualified Scientist, Adult Sponsor, or judge.
价格与内测Pricing and beta
现在是免费试点The first round is a free pilot
学生赛季版首轮免费：我们把整套交付做完给你，你先看交付内容，觉得有用再谈价格。正式价格在试点之后公布，报价逐单回信确认。The first round of the Student tier is free: you get the full delivery and judge the work first. Formal pricing follows the pilot, confirmed per order by email.
交付时限也逐单在回信里约定。事前审批的时序（例如人类被试研究须在招募前拿到 IRB 裁定）是官方规则的硬约束，不是我们能替你压缩的。Turnaround is likewise agreed per order by email. Prior-approval timing (for example, an IRB determination before recruitment in a human-participant study) is an official hard constraint — not something we can compress for you.
申请 学生赛季版（封闭内测）Request the Student tier (closed beta)
不想写邮件的话，填几项基本信息就行，我们会主动联系你。No email needed — a few basics and we’ll reach out to you.
FAQ
常见问题Frequently asked
买了 学生赛季版就能晋级 ISEF 吗？Does the Student tier get my project to ISEF?
否。任何档位都不保证、不预测晋级或获奖，概率与名次预测也不提供。交付的是缺口与时序风险的证据清单和整改行动表；评审由关联赛与 ISEF 的评委决定，与本产品无关。No. No tier guarantees or predicts advancement or awards, and none outputs probabilities or placement forecasts. What you get is an evidence-based list of gaps and timing risks plus a remediation table; judging is decided by the affiliated fair and ISEF judges and is unrelated to this product.
我的实验已经开始了，还没拿到 IRB 裁定，你们能补吗？My experiment already started and I have no IRB determination yet — can you fix that?
我们不能，也不会替你断言"没问题"。规则要求非豁免的人类被试研究在招募、接触、干预或采集数据之前取得责任 IRB 的豁免或批准裁定；已采集数据能否使用，由该 IRB 裁定。我们能做的是：把时序事实如实写清、指出责任机构是谁、把问询与补件排进倒排表，并提醒在裁定到位前停止新的采集。We cannot, and we will not assert on your behalf that it is fine. The rules require a non-exempt human-participant study to obtain the responsible IRB's exemption or approval determination before recruitment, interaction, intervention, or data collection; whether already-collected data may be used is that IRB's call. What we can do is state the timing facts plainly, identify the responsible body, schedule the inquiry and the missing documents into the countdown, and remind you to stop further collection until the determination is in hand.
我用过 AI，会不会有问题？I used AI — is that a problem?
用没用过不是判定线，是否写了初稿、是否披露才是。官方规则允许把生成式 AI 作为被引用的项目资源，但它不得撰写研究计划、摘要、海报或引文，呈现材料必须是学生自己的语言；披露在 Form 2A 上并附 prompt-and-response 日志。我们会核对披露是否与实际使用对应，并提醒缺失时段如实说明。不补造记录，也不代写替换文本。Whether you used it is not the test — whether it wrote the initial draft and whether it is disclosed is. The official rules allow generative AI as a cited project resource, but it may not write the research plan, abstract, poster, or citations, and presented material must be in the student's own words; disclosure goes on Form 2A with the prompt-and-response log attached. We check that the disclosure matches actual use and remind you to state missing periods honestly — we never fabricate a log, and we never ghostwrite replacement text.
教练或机构可以代学生对接吗？Can a coach or institution handle this on the student's behalf?
可以。ISEF 没有禁止机构指导参赛项目的规则，教练、工作室与学校都是我们的正常客户，多项目请直接看 机构学校版。唯一的边界在"谁写初稿"：研究计划、摘要与海报文本必须是学生自己写的，这一条对家长、教练、机构和我们同样适用。Yes. ISEF has no rule prohibiting institutional mentoring, and coaches, studios, and schools are normal customers — for multiple projects see the Institution tier. The only boundary is who writes the initial draft: the research plan, abstract, and poster text must be the student's own, and that applies equally to parents, coaches, institutions, and to us.
我的材料会被怎么处理？How are my materials handled?
仅在本地处理，用于生成本次交付；不用于训练模型，不对外分享。我们只收状态与非敏感证据指针，不收签名扫描件、知情同意书正文、伦理审查原件与参与者名单；学生姓名、出生日期、联系方式、参与者身份与任何医疗标识符不进入台账、报告正文或文件名。交付后的保留 / 删除方式逐单约定（正式保留政策尚未定稿）。Processed locally only, solely to produce your delivery; never used to train models, never shared. We collect status and non-sensitive evidence pointers only — never signature scans, consent text, original ethics documents, or participant lists — and student names, dates of birth, contact details, participant identities, and any medical identifier never enter the ledger, report bodies, or filenames. Post-delivery retention or deletion is agreed per order (the formal retention policy is not yet final).