Thoth

io.github.ahmedEid1由 GitHub 账号 ahmedEid1 发布v1.0.1更新于 Oct 24, 2026

Read-only MCP over an agentic SLR workspace with per-claim citation verification

已验证Streamable HTTP可网页运行网页搜索与抓取知识与记忆

概览

AI 生成的概览

以只读方式访问 Thoth 的智能体系统性文献综述工作区,并提供逐条论断的引文核验报告。

可以做什么
  • 列出你的系统性文献综述及其评审分数与引文忠实度
  • 读取已完成综述的 Markdown 草稿
  • 获取逐条论断的引文审计,标注每条引用为有据、无据或存疑
  • 查看发现阶段检索到的论文及其抓取与筛选状态(v2)
  • 查看发现器生成的检索式及各数据源的错误信息(v2)
适用场景
当你已经自行部署 Thoth,并希望助手调取已完成的文献综述、草稿和引文审计,或在阅读草稿前确认其引用是否确实与原文相符时使用。
运行要求
需要一个运行中的 Thoth 部署:托管应用已下线,需自行部署(Node.js/pnpm、用 Docker Compose 启动 Postgres、MinIO 与 Langfuse、Prisma 迁移、Trigger.dev worker),并把客户端指向自己的 MCP 地址。OAuth 2.1 + PKCE 与动态客户端注册通过 Clerk 在浏览器中完成。自托管需要 Clerk 与 Trigger.dev 密钥以及某个 LLM 提供商的密钥,例如 MISTRAL_API_KEY。
安装前请注意
公开托管端点已停用,注册表中的地址无法直接使用,必须自行部署并保护实例。工具为只读且限定在本人账户,但会暴露综述草稿、引文审计和发现的论文,审计日志会保存每次调用输入的 SHA-256 哈希。自托管需配置第三方凭据(Clerk、Trigger.dev、LLM 提供商),并会向外调用 OpenAlex、arXiv 和 Exa。README 显示论断忠实度仅 38%,v2 的发现与筛选仍在校准,输出应视为待核实的草稿。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Thoth,将其添加到工作区。
  2. 连接 所需的 凭据。它们会被私密存储,绝不会在公开页面上展示。
  3. 为需要使用其工具的对话启用该服务。

Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体,或 SourceWeft 对话中。智能体会读取该服务器的安装指南,先向你展示发布者、端点和登录方式,经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/io.github.ahmedEid1%2Fthoth/install.md,并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前,先告诉我它的发布者、端点和登录方式,等我确认。不要把 API 密钥写进命令或我会提交的文件;问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。 客户端会在首次连接时完成 OAuth 登录。

{
  "mcpServers": {
    "thoth": {
      "type": "http",
      "url": "https://thoth-slr.vercel.app/api/mcp/mcp"
    }
  }
}

README

[Thoth — sacred ibis logo]

Thoth

Agentic systematic literature reviews — with every citation checked against the source.

Named for Thoth, ancient Egypt's ibis-headed god of writing and scribes.

[Hosted app] [Tests] [App release] [Deploy cost] [License: MIT]

The hosted app ran live in 2026 and is now retired. The screenshots and walkthroughs below are from it; you can still run it yourself.

[Browsing a completed Thoth review — draft, critic score, and per-claim citation audit]

What is Thoth?

Systematic literature reviews are slow to write — and when you ask an LLM to write one, it confidently invents citations and statistics that aren't in any paper.

Thoth does both halves and checks its own work. Give it a research question and it discovers relevant papers, reads them, drafts an evidence-grounded review — then runs a verification pass (cite_check) that compares every cited claim against the source paper and flags anything unsupported before you read the draft. The result is a review with a critic score, a citation-faithfulness percentage, and a per-claim audit you can trust.

It ran (hosted, in 2026) as a polished web app, a public eval dashboard, and an authenticated MCP server your AI assistant could call directly; it still runs on your own machine or host.

See it work

Claude.ai catches 6 fabricated citations in a real draft — using Thoth's audit:

Connected to Thoth via the official MCP Registry, Claude calls get_citation_audit on one deliberately-weak review (faithfulness 0.13 for that single review) and identifies all 6 unsupported claims — every one citing the same paper, with invented percentages that aren't in the source. This is cite_check doing its job: it's a single-review audit sample, not the golden-set aggregate (the /evals dashboard had that).

Every claim, scored against its source — the /showcase review (no login needed). The figures on this card (critic 4.2/5, faithfulness 75%, 8/8 citations checked, 2 unsupported) are this one review's scores — a worked example, not the aggregate:

Evaluated in public — the /evals dashboard tracked inclusion recall / precision, claim faithfulness and coverage over an 18-question versioned golden set (7 of 18 have data at the latest sweep's commit), regenerated in CI and published with the last-run date (last run 6 July 2026, before the hosted app was retired), so a regression was a public, falsifiable signal:

You approve every step — three human-in-the-loop gates (review plan → review discovered papers → approve included papers); nothing runs unattended:

Key features

  • 🔎 cite_check — verifiable citations. Every [paper_id] in the draft is scored against the cited paper and labelled supported / unsupported / unclear, so the LLM can't quietly hallucinate a citation. The verdict report is published per claim, not summarised away (claim faithfulness on the golden set: 38%). This is the core differentiator: the citations are measured, not asserted. Paper selection is scored separately: the papers Thoth includes are compared with each golden question's expected paper list — inclusion precision 97%, recall 74%, averaged over the 9 golden questions that have an inclusion score. That measures which papers it picks, not whether its citations are correct, and precision is 1.0 by construction on goldens with no distractor papers (see known limitations).
  • 🌐 Outbound web search (v2 — under active evaluation). An outbound discoverer → fetcher → screener path is wired across OpenAlex, arXiv, and Exa: it fetches open-access PDFs, OCRs them, and screens each against your plan, so you can run uploaded-only, hybrid, or fully autonomous discovery. The discovery and screening axes are v2 and still being calibrated — they're tracked openly on /evals (both at 0% at the last run) rather than shipped as a silent claim.
  • 🔌 Authenticated, registered MCP server. OAuth 2.1 + PKCE + Dynamic Client Registration via Clerk, SHA-256 audit logging, rate limits — listed in the official MCP Registry. Most public MCP servers ship with no auth; this one doesn't.
  • 📊 Public eval dashboard. Inclusion recall / precision, faithfulness and coverage over a versioned golden set, regenerated in CI and stamped with the last-run date, rendered at /evals — an eval regression was a public signal, not a hidden one. The weekly eval run is retired; the last run was 6 July 2026.
  • 💸 6 LLM providers, $0/mo by default. Swap providers with one env var; the hosted app ran on the Mistral free tier, and the entire stack deploys on free tiers for $0/mo.

🚀 Quickstart

The hosted app ran live in 2026 and is retired, so run your own instance (below).

Connect it to your AI assistant — paste your deployment's MCP URL into claude.ai (Pro/Max), Claude Desktop, Cursor, or any MCP client (OAuth runs in your browser; no token to copy):

https://<your-host>/api/mcp/mcp
Read-only MCP tools (scoped to your account)
  • list_reviews — your reviews with critic + faithfulness scores
  • get_review_draft — the markdown draft of a completed review
  • get_citation_audit — the per-claim cite_check verdict report
  • list_discovered_papers (v2) — papers the discoverer surfaced, with fetch + screening status
  • get_search_queries (v2) — the queries the discoverer generated + per-provider errors

Full reference: docs/mcp/tools.md · auth + audit model: docs/mcp/security.md

[Adding Thoth as a custom MCP connector in claude.ai — paste the URL, OAuth via Clerk + Dynamic Client Registration]
Adding Thoth as a custom connector in claude.ai — OAuth runs in your browser (Clerk + DCR), no token to copy.

Run it locally:

bash
git clone https://github.com/ahmedEid1/thoth.git && cd thothcp .env.example .env        # Clerk + Trigger.dev keys + MISTRAL_API_KEYdocker compose up -d        # postgres, minio, langfusepnpm install && pnpm prisma migrate devpnpm dev                    # Next.js on :3000pnpm dev:trigger            # Trigger.dev worker (separate terminal)

Full setup, the agent pipeline, and the v2 flow: docs/architecture.md.

Proof

Hosted appRan live in 2026 (Clerk sign-in, sample review at /showcase); retired on 1 Nov 2026
Public evals/evals (last run 6 July 2026) — inclusion precision 97%, recall 74% (papers included vs. each question's expected list, averaged over the 9 golden questions with an inclusion score) on a versioned 18-question golden set (7 of 18 have data at the latest sweep's commit; faithfulness 38% / coverage 32% tracked in the open as the set fills out; discovery/screening v2 under calibration). Regenerated in CI, published with the last-run date — a regression is a public signal.
MCP Registryio.github.ahmedEid1/thoth — listed; its remote endpoint was retired with the hosted app
Tests676 unit/integration + 22 live e2e against the deployed instance (MCP transport, real-browser, authenticated walkthroughs, full agent runs) — all green; tsc + lint clean
Audit logEvery MCP call recorded with a SHA-256 input hash; no raw input stored
Deploy cost$0/mo — Vercel + Neon + Cloudflare R2 + Langfuse + Trigger.dev, all free tiers (self-host option)

For engineers

Thoth is a LangGraph StateGraph driven by a Trigger.dev worker, with durable human-in-the-loop gates, a per-run cost cap, and exactly-once gate delivery. Next.js 16

  • TypeScript (strict), Postgres + Prisma, Clerk auth (web + OAuth 2.1 for MCP), S3-compatible storage, Mistral OCR, Langfuse tracing.

Credits

Ibis icon by Delapouite under CC BY 3.0, via game-icons.net.

License

MIT © 2026 Ahmed Hobeishy

来源:README.md,提交 d7c5b4b

工具

0
该服务器需要登录,因此 SourceWeft 无法在没有账号的情况下读取其工具,也未收录。已连接账号所能看到的内容可能因账号及其权限而异。

版本历史

1
  1. v1.0.1最新Sep 16, 2026

评分

已满足适用检查项的 60%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分

    请保持端点能够响应 MCP 握手;SourceWeft 每 30 天探测一次。

  • 已声明认证方式必需10 分
  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list,或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分