Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@plaud-mcp get the transcript of my latest recording"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server for Plaud transcripts via CDP proxy through the running Plaud Desktop app.
Claude Code → MCP Server → CDP (WebSocket) → Plaud Desktop → Plaud APIThe MCP connects to the running Plaud Desktop Electron app via Chrome DevTools Protocol:
Sends SIGUSR1 to the Plaud Desktop process to enable Node.js inspector (port 9229)
Connects via WebSocket to the inspector
Executes API calls through the app's own authenticated $fetch function
Returns results back through MCP tools
No token extraction, no cookies, no API keys. Uses the app's live authenticated session directly.
Related MCP server: Plaud Notes MCP Server
Plaud Desktop - installed and signed in
Python 3.10+
uv (recommended)
# Install as a CLI tool (recommended)
uv tool install "plaud-mcp @ git+https://github.com/davidlinjiahao/plaud-mcp"
# Or install from local source
uv tool install --force "plaud-mcp @ ."Add to ~/.claude.json:
{
"mcpServers": {
"plaud": {
"command": "${HOME}/.local/bin/plaud-mcp",
"args": [],
"env": {}
}
}
}No API keys or tokens needed. Just ensure Plaud Desktop is running.
Tool | Description |
| Verify Plaud Desktop is available |
| Total number of recordings |
| Files from the last N days |
| Files with optional date filters |
| Metadata for a specific file |
| Full transcript with speaker labels |
| AI-generated summary |
| Search transcripts by content |
Launch the Plaud Desktop app and sign in.
The SIGUSR1 signal may have failed. Ensure Plaud Desktop is the main process, not a helper.
search_transcripts fetches and searches client-side. Reduce the days parameter.
Plaud's API validates auth at the Chromium network stack level - tokens extracted from LevelDB don't work with standard HTTP clients (httpx, curl, curl_cffi with Chrome impersonation all return 401). The CDP approach bypasses this entirely by executing requests through the app's own authenticated context.
uv venv && source .venv/bin/activate
uv pip install -e ".[dev]"
ruff check src/ && pyright src/Check if Plaud Desktop is available and authenticated.
Inputs: no named parameters. Returns: object
Get metadata for a specific Plaud file.
Inputs: file_id (required). Returns: object
Get the total number of Plaud files.
Inputs: no named parameters. Returns: object
Get Plaud files with optional date filters (ISO format, e.g. '2024-01-01').
Inputs: start_date, end_date, limit. Returns: result
Get Plaud files from the last N days.
Inputs: days. Returns: result
Get AI-generated summary for a Plaud file.
Inputs: file_id (required). Returns: object
Get full transcript with speaker labels for a Plaud file.
Inputs: file_id (required). Returns: object
Search recent transcripts for matching content. Searches client-side.
Inputs: query (required), days. Returns: result
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 8 tools
Most tools have distinct purposes: connection check, file retrieval variants, summary, transcript, and search. However, 'get_files' and 'get_recent_files' could cause slight confusion as both retrieve files but with different filtering.
All tool names follow a consistent verb_noun pattern using snake_case, such as 'check_connection', 'get_file', 'search_transcripts'. The naming is predictable and easy to understand.
With 8 tools, the server is well-scoped for its purpose of interacting with Plaud files. Each tool provides distinct functionality without redundancy or excessive complexity.
The tool set covers retrieval (files, summaries, transcripts) and search, but lacks any create, update, or delete operations. This leaves a notable gap for file management, though it may be intentional due to API limitations.
Your Plaud recordings in natural language: list recordings, read speaker-attributed transcripts and
Fathom MCP server for ChatGPT, Claude and any AI agent: list calls, complete action item, add action item, share call, set link access, regenerate summary, update settings and delete call in your own Fathom account. Cotlet works in the account through your Cloud Browser, a real Chrome where you sign in to Fathom yourself; Cotlet does not store your password and you need no Fathom API key. About the official Fathom API: Fathom has a REST API and an official MCP server, but both only read your calls (checked 22 Oct 2026). The API lists meetings and returns summaries, transcripts and action items, requests recording downloads and manages webhooks. Setup and FAQ: https://cotlet.dev/integrations/fathom-video-api-mcp-server
Otter.ai MCP server for ChatGPT, Claude and any AI agent: search transcripts, export transcript, share conversation, add highlight, upload audio and list action items in your own Otter.ai account. Cotlet works in the account through your Cloud Browser, a real Chrome where you sign in to Otter.ai yourself; Cotlet does not store your password and you need no Otter.ai API key. About the official Otter.ai API: Yes, but narrowly (checked 16 Oct 2026). Otter.ai has an official hosted MCP server with three tools (user info, search, fetch) that only read, and a public REST API available to Enterprise workspaces that reads conversations, transcripts, action items, insights, outlines, audio, channels and the workspace, and imports a recording from a file URL. Setup and FAQ: https://cotlet.dev/integrations/otter-api-mcp-server
Bluedot MCP server for ChatGPT, Claude and any AI agent: list meetings, search transcripts, get meeting summary and transcript, move meeting to collection, create collection, create notes template, apply notes template and add comment in your own Bluedot account. Cotlet works in the account through your Cloud Browser, a real Chrome where you sign in to Bluedot yourself; Cotlet does not store your password and you need no Bluedot API key. About the official Bluedot API: Bluedot documents no REST API, but it has an official MCP server and webhooks (checked 20 Oct 2026). The MCP server has 8 tools that only read meetings, transcripts, summaries, collections and participants, and the pricing page lists it from the Basic plan. Setup and FAQ: https://cotlet.dev/integrations/bluedot-api-mcp-server
