MCP

Connect your AI to AudioMap

In 3 steps, your favorite AI can search, read and summarize your transcribed voice notes. Works in Claude, ChatGPT, Claude Code, Cursor, and any MCP-compatible tool.

Go to MCP dashboardSee per-platform guides

Three ways to use it

From trying it without signing up to full access with your personal library.

No signup

Public demo

Try AudioMap MCP without an account. Demo dataset, rate-limited.

  • ✓No signup, no token
  • ✓10 requests per IP every 10 min
  • ✓Limited read tools
  • ✓Ideal to evaluate before signing up
https://api.audiomap.ai/mcp/demo
€0 · Free

Free account

Your own notes library. Full read, write within plan credits.

  • ✓120 transcription minutes / month
  • ✓Full read tools (free)
  • ✓Write tools using your credits
  • ✓1 active MCP session
Create free account →
From €19/mo

Pro / Power

Concurrent MCP sessions, generous credits, processing priority.

  • ✓1,800 minutes / month (Pro)
  • ✓Up to 5 concurrent MCP sessions
  • ✓No overage: 402 when exhausted
  • ✓Priority support
See plans →

Quick start — 3 steps

1

Create a persistent API key at /dashboard/api-keys (recommended) or a session token at /dashboard/mcp.

2

Copy the generated "connection prompt" for your AI. Works in Claude, ChatGPT, Claude Code and more.

3

Paste it into your AI. It configures itself and shows you what to ask.

Generate my connection prompt

Per-platform guides

Specific instructions for each AI

🤖Claude.aiEasy+
  1. 1Open Claude.ai and go to Settings
  2. 2Click "Connectors"
  3. 3Click "Add custom connector"
  4. 4Paste the URL: https://api.audiomap.ai/mcp/stream
  5. 5Enter your am_live_* Bearer token when prompted
  6. 6Save — done!
💬ChatGPTPaid plans only+
Requires ChatGPT Plus, Pro, Team, or Enterprise. The free plan does not support MCP servers.
  1. 1Open ChatGPT and go to Settings (gear icon)
  2. 2Go to Apps & Connectors
  3. 3Enable "Developer Mode"
  4. 4Click "Create connector"
  5. 5Name it "AudioMap"
  6. 6Paste the server URL: https://api.audiomap.ai/mcp/stream
  7. 7Select Bearer token authentication and paste your am_live_* token
⌨️Claude CodeCLI+
Claude Code does NOT read mcpServers from settings.json. Use the claude mcp add command (recommended) or create .mcp.json at the project root.
  1. 1Open your terminal in any project where you use Claude Code
  2. 2Run: claude mcp add --transport http -s user audiomap <URL> --header "Authorization: Bearer <TOKEN>"
  3. 3Replace <URL> and <TOKEN> with values from your dashboard
  4. 4Restart the Claude Code session
  5. 5Ask: "list my latest audiomap notes"

Recommended command (whole machine)

claude mcp add --transport http -s user audiomap https://api.audiomap.ai/mcp/stream --header "Authorization: Bearer <YOUR_TOKEN>"

Alternative: .mcp.json in the project

{
  "mcpServers": {
    "audiomap": {
      "type": "http",
      "url": "https://api.audiomap.ai/mcp/stream",
      "headers": {
        "Authorization": "Bearer <YOUR_TOKEN>"
      }
    }
  }
}
🛠️Claude Desktop, Cursor, VS CodeAdvanced+
These tools use a JSON configuration file. Paste the mcpServers block in the corresponding path.

Paths:

  • Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Cursor: .cursor/mcp.json
  • VS Code: .vscode/mcp.json
  • Claude Code: .mcp.json (or use claude mcp add)
{
  "mcpServers": {
    "audiomap": {
      "type": "http",
      "url": "https://api.audiomap.ai/mcp/stream",
      "headers": {
        "Authorization": "Bearer <YOUR_TOKEN>"
      }
    }
  }
}
✨GeminiNot supported+
Gemini web/app does not support connecting custom MCP servers. If you use Gemini CLI you can do it programmatically; see its documentation.

Available tools

14 tools total. Read tools are free; write tools consume credits.

search_notes

Search notes by text (title, description, transcript).

list_recent_notes

Last N notes with metadata.

get_note

Full note: transcript + AI summary + highlights.

get_transcript_segments

Segments with timestamps and speakers.

list_folders

Folders with note count.

list_highlights

Highlighted fragments of a note.

get_outline

Auto-generated chapters with timestamps.

upload_audio_url

Upload audio from public URL. 50 cred/min.

wait_for_transcription

Polling with timeout until status ready.

generate_ai_notes

Templates: meeting, sales, legal, medical… 100 cred.

generate_outline

Outline/chapters with AI. 50 cred.

replace_in_transcript

Literal search & replace over the transcript. Free.

ask_note

Ask about a note (RAG + timestamps). 10 cred.

extract_clip

Extract MP3 clip from a range. 1h presigned URL. 5 cred.

Full descriptions and annotations: https://api.audiomap.ai/mcp/info

Ready to connect your AI?

Generate your personalized connection prompt in one click. Your AI does the rest.

Go to MCP dashboardCreate persistent API key
Connect your AI to AudioMap | MCP setup guide · AudioMap