Transcribe by STIV Labs
MCP Setup

Give this page to the agent and it will connect MCP on its own

The agent can use this page to understand the connection steps. You only need to create an MCP token in your account and share this setup page.

Language:English·Russian
Back to account

Available MCP commands

get_usage_summary

Shows the current minute balance and usage summary for this account.

list_recent_transcriptions

Returns recent transcription jobs with status and short text preview.

get_job

Returns the current status and result payload for a specific transcription job.

transcribe_file

Uploads a local file for transcription and creates a new job.

Open full command reference

Current connection mode

Right now each client can run the MCP adapter locally and connect it to the platform with a personal MCP token.

Step 1. Create an MCP token

Open your account page, create a dedicated MCP token, and copy it immediately. The full token is shown only once.

Step 2. Start the MCP adapter

Run the stdio MCP adapter and pass your token plus the API base URL. If you launch it from a monorepo checkout, also pass --directory <project-root> so uv can resolve the correct workspace.

Step 3. Add it to your AI client

Use any stdio-compatible MCP client. For Claude Code, prefer `claude mcp add ...` instead of editing `~/.claude/settings.json` by hand. Keep the token in environment variables instead of hardcoding it into prompts.

Local launch command

UV_CACHE_DIR=/tmp/uv-cache uv run --directory <PROJECT_ROOT> python -m transcription_mcp.server \
  --api-base-url https://api.aiasr.ru \
  --mcp-token <YOUR_MCP_TOKEN>

Example MCP client config

This JSON block is a generic stdio MCP example. For Claude Code, use `claude mcp add` so the config is written to `~/.claude.json` in the supported format.

{
  "mcpServers": {
    "audio-transcription-workflow": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "<PROJECT_ROOT>",
        "python",
        "-m",
        "transcription_mcp.server",
        "--api-base-url",
        "https://api.aiasr.ru"
      ],
      "env": {
        "UV_CACHE_DIR": "/tmp/uv-cache",
        "TRANSCRIPTION_MCP_TOKEN": "<YOUR_MCP_TOKEN>"
      }
    }
  }
}