QuillHub

ai.quillhub由 quillhub.ai 發布官方發布者v1.0.1更新於 Oct 27, 2026

Meeting transcripts for AI agents: search calls, read who said what, transcribe files and links.

已驗證Streamable HTTP可網頁執行效率與工作流知識與記憶溝通與協作

概覽

AI 產生的概覽

託管式 MCP 伺服器,讓 AI 助理檢索 QuillHub 會議逐字稿、發言者原話、決議與待辦事項。

可以做什麼
  • 跨工作區與專案搜尋、列出錄音,含摘要、決議、待辦事項與與會者
  • 以摘要、全文、發言片段、對話、段落、章節或字幕讀取單一錄音
  • 依日期與專案篩選,取出某人在各場會議中帶時間戳的原話
  • 瀏覽追蹤對象(人物、交易、專案、客戶),查看動態筆記、未結承諾與修訂時間軸
  • 列出跨會議發現的矛盾與待決問題,讀取 AI 維護的專案簡報
  • 轉錄連結或內嵌檔案,並等待或取消轉錄工作
適用情境
當會議錄音存放在 QuillHub,且希望助理在對話中回答過去通話的問題、引用某人原話、會前準備簡報,或轉錄新的檔案與連結時適用。
執行需求
遠端託管端點,使用 Streamable HTTP,無需安裝或自行架設。必須驗證:互動式用戶端採 OAuth 2.1 加 PKCE,無介面用戶端則以 Authorization Bearer 標頭傳 API 金鑰(通常經 mcp-remote 橋接)。需要 QuillHub 帳號。
安裝前請注意
助理可讀取整個會議檔案,包括同事共享的錄音,應謹慎控管授權範圍。轉錄會消耗 QuillHub 餘額(註冊贈送 60 分鐘,之後依方案計費)。建立轉錄會把連結或上傳檔案傳送給 QuillHub。不要把 API 金鑰提交進設定檔,並在後台定期輪換。

安裝

在 SourceWeft 中

  1. 開啟 儀表板中的 QuillHub,將其新增到工作區。
  2. 連接 所需的 憑證。它們會被私密儲存,絕不會在公開頁面上展示。
  3. 為需要使用其工具的對話啟用該服務。

Web executable,透過 Streamable HTTP。 遠端服務在工作區中設定後即可從網頁執行環境執行。

交給代理程式安裝

把這段提示詞貼上到 Claude Code、Codex、Cursor 等代理程式,或 SourceWeft 對話中。代理程式會讀取該伺服器的安裝指南,先向你展示發布者、端點和登入方式,經你同意後再把它新增到你的客戶端。金鑰不會寫進任何命令。

閱讀 https://sourceweft.com/mcp/ai.quillhub%2Fmcp/install.md,並按其中說明把這個 MCP 伺服器接入我的客戶端。修改任何設定之前,先告訴我它的發布者、端點和登入方式,等我確認。不要把 API 金鑰寫進命令或我會提交的檔案;問我金鑰應從哪裡取得。

檢視代理程式所依據的安裝指南

其他 MCP 客戶端

把它新增到你客戶端的 mcpServers 設定中。 客戶端會在首次連接時完成 OAuth 登入。

{
  "mcpServers": {
    "quillhub": {
      "type": "http",
      "url": "https://mcp.quillhub.ai/mcp"
    }
  }
}

README

[QuillHub logo]

QuillHub MCP Server

Your meeting transcripts, readable by Claude, Cursor and any other MCP client.
Search past calls, read who said what, pull decisions and action items, transcribe new files and links.

Setup guide · Docs · quillhub.ai · Desktop app


Endpoint: https://mcp.quillhub.ai/mcp (Streamable HTTP, hosted) Auth: OAuth 2.1 with PKCE (interactive clients) or API key qai_live_... (scripts, CI) Source: this repository documents the hosted server; the server itself is operated by QuillHub. There is nothing to install or self-host.

What it is

QuillHub records meetings from your desktop without a bot in the call (Zoom, Meet, Teams, phone, in person) and transcribes uploaded files and YouTube links in 98+ languages. Every recording comes back split by speaker with timestamps, chapters, a summary, decisions and action items.

The MCP server gives your AI assistant direct access to that archive. Ask Claude "what did we agree with the client last Tuesday", "what has Anna said about pricing this month", or "transcribe this file and draft the follow-up", and it calls the right tools itself.

Quick start

Step-by-step guide with screenshots: QuillHub + Claude/Cursor via MCP.

Claude Desktop, Cursor, VS Code and other OAuth-capable clients

Point the client at the URL with no Authorization header. The client opens a browser, you sign in to QuillHub, done.

json
{  "mcpServers": {    "quillhub": {      "url": "https://mcp.quillhub.ai/mcp"    }  }}

Claude.ai / Claude Desktop: Settings → Connectors → Add custom connector, paste https://mcp.quillhub.ai/mcp.

Claude Code:

bash
claude mcp add --transport http quillhub https://mcp.quillhub.ai/mcp

API key (headless, CI, clients without OAuth)

Get a key in the Developers dashboard (issued automatically at sign-up), then use the mcp-remote bridge:

json
{  "mcpServers": {    "quillhub": {      "command": "npx",      "args": [        "-y",        "mcp-remote",        "https://mcp.quillhub.ai/mcp",        "--header",        "Authorization:Bearer qai_live_YOUR_KEY"      ]    }  }}

Write Authorization:Bearer ... with no space after the colon: some mcp-remote versions split arguments on spaces.

Tools

ToolWhat it does
search_meeting_transcriptsFull-text search across meetings in a workspace: titles, summaries, decisions, action items and transcript text. Morphological matching in English and Russian.
list_transcriptionsPaginated list of recordings with summary, key theses, decisions, action items, participants, project. Filter by text, date range, workspace, project.
get_transcriptionOne recording in the format you need: summary, text, segments, dialog (who said what), paragraphs, chapters, subtitles. Supports time windows for long calls.
get_person_speechTimestamped quotes from one person across every meeting, with filters by date and project.
find_subjectsFind tracked subjects (people, deals, projects, clients, candidates, custom types) with their AI-maintained note, open commitments and mention counts.
get_subject_pageFull page for one subject: living note, open commitments, every meeting mention with quotes and speakers, unresolved questions.
get_state_timelineHow a subject's note changed over time: revisions with diffs and the recording that caused each change.
list_open_questionsContradictions and ambiguities the AI found across meetings (e.g. two different budget figures) that need a human decision.
get_project_memoryAI-maintained briefing for a project or workspace: purpose, glossary, key people, activity digest.
list_subject_typesSubject types available in a project and the fields each one tracks.
list_workspacesYour workspaces and the projects inside them, so the agent can read teammates' shared recordings.
create_transcriptionTranscribe a URL (YouTube, direct file link) or an inline base64 file up to 25 MB.
wait_for_transcriptionWait for a job to finish (up to 300 s per call), no polling loop needed.
cancel_transcriptionCancel a queued or running job.
get_accountAccount, balance and subscription. Cheap liveness check.
get_developer_docsQuillHub developer docs as markdown, for agents writing integration code.

Example prompts

  • "Find every call where we discussed the migration deadline and list the decisions."
  • "What did Maria commit to in last week's meetings? Quote her."
  • "Transcribe https://youtu.be/... and give me the chapters with timestamps."
  • "Brief me on the Acme project before tomorrow's call."
  • "When did we first hear about the budget cut, and what changed after the last meeting?"

Limits and pricing

  • Connecting MCP is free. Transcription uses your normal QuillHub balance: 60 free minutes on sign-up, no card. Plans at quillhub.ai/en/pricing.
  • Inline base64 uploads are capped at 25 MB; for larger files pass a URL. Files up to 10 hours / 5 GB via the app.
  • No webhooks over MCP. Use the REST API webhooks for push delivery.

Security and privacy

  • The server is stateless and forwards your token to the QuillHub REST API; it stores no recordings and no secrets.
  • Recordings are encrypted at rest and in transit, never used to train models, and can be deleted at any time.
  • Don't commit real API keys into config files. Rotate keys in the Developers dashboard.

Kiro power

This repository is also a Kiro power: plugin.json, mcp.json and POWER.md sit at the root.

Also available

Support

Open an issue in this repository or write to [email protected].

Privacy policy: quillhub.ai/en/legal/privacy

QuillHub is built by Axevia Labs LLC.

License

The MIT license covers the files in this repository (documentation and config examples). The hosted server at mcp.quillhub.ai is a proprietary service.

來源:README.md,提交 204dc46

工具

0
該伺服器需要登入,因此 SourceWeft 無法在沒有帳號的情況下讀取其工具,也未收錄。已連接帳號所能看到的內容可能因帳號及其權限而異。

版本歷史

1
  1. v1.0.1最新Sep 23, 2026

評分

已滿足適用檢查項的 85%

該評分衡量此收錄條目的完整程度,依據註冊中心條目、端點與公開來源計算得出。它不是對伺服器本身的評價,也不表示伺服器安全或目前可用:可連通性是帶日期的檢查結果,評分也不包含安裝量或使用者評價。

  • 近 30 天內可連通必需25 分
  • 已宣告認證方式必需10 分
  • 已列出工具15 分

    請讓伺服器在無憑證時回應 tools/list,或發布可在沙箱中無憑證啟動的軟體包。

  • 描述必需15 分
  • 圖示10 分
  • 文件15 分
  • 授權條款10 分