Transcription for your scripts and AI apps

Use Parlova through its REST API, command-line tool or local MCP server. Plus and Pro include developer access; submissions use the same account allowance as the web app.

What do I need?

You need a Plus or Pro account. You can try transcription quality on the free web plan before upgrading. The Python CLI requires Python 3.8 or newer and uses the standard library.

Install and sign in

Install the published Parlova package, then sign in through the browser approval flow:

pip install --user parlova
parlova login

Check the displayed sign-in code and the requesting app before approving access. Your password stays in the web sign-in flow; do not paste it into an assistant or configuration file.

Transcribe from the command line

parlova transcribe meeting.m4a --speakers
parlova list
parlova get <job-id> --format srt -o meeting.srt

Supply the path to a local recording or an accessible YouTube URL. File duration, upload size, queue and remaining-allowance limits still apply. Formats and speaker options depend on your plan.

Connect Claude, Cursor or VS Code with MCP

The MCP server runs locally and sends transcription requests to your Parlova account. Generate the configuration for your client:

parlova mcp-config
parlova mcp-config --app cursor
parlova mcp-config --app vscode
parlova mcp-config --app claude-code

Follow the generated configuration instructions and restart your client where required. Generated absolute paths avoid problems when desktop apps do not inherit your terminal's PATH. The MCP server command is parlova mcp.

Available tools include transcribe_file, transcribe_youtube, get_transcript, list_transcripts and get_usage. For example, ask your assistant to transcribe a local interview, retrieve its transcript, then identify decisions or action items. Any interpretation by that assistant is separate from the original transcription and should be checked.

Use the REST API

Sign in and open Developers → API tokens to create a token. Use it as a Bearer token in the Authorization header. The developer section in the app includes request examples; the public CLI and MCP source also shows the integration.

Keep tokens out of browser code, public repositories and chat transcripts. Revoke tokens you no longer need. The account's owner is responsible for authorised access to recordings and transcripts.

What data does the integration use?

A file submitted for transcription is uploaded to Parlova; a YouTube request supplies the video link. When an assistant retrieves a transcript, its own service may process that text under its own policies. Review both Parlova's Privacy Policy and your assistant's data controls before using confidential material.

Where can I get help?

See the installation and tool reference for additional configuration examples. For account or billing help, contact jurie24102@gmail.com.

Built and operated by Jurie Wagener, trading as SoleCreations, an independent developer in South Africa. Contact Jurie.