Model Context Protocol

Video narration from your IDE

Narrate, transcribe, dub, translate, text-to-speech, and clone voices — all from Cursor, VS Code, Claude, or any MCP-compatible client. One API key, one line of config.

Works with any MCP client · Supports 20+ languages · AI voices & voice cloning

See setup

What it looks like

"Narrate this demo video with the male1 voice: https://storage.example.com/demo.mp4"

I'll narrate your video using the male1 voice. Processing now...

narrate_video_full(video_source="...", voice_type="male1")

Done! Here's your narrated video: download link

What you can do

Narrate

Silent video → AI script → narrated video with voiceover. Choose an AI voice or clone your own.

Synced Transcript

Silent video → AI watches and writes a natural-language script → timed segments synced to the video.

Transcribe

Video with existing voice → speech-to-text → timed transcript with segments.

Translate

Translate transcripts across 20+ languages. New uploads or existing videos in your library.

Dub

Full auto-dubbing: transcribe → translate → TTS with cloned speaker voice → dubbed video.

Text-to-speech

Turn any text into spoken audio with AI voices or a cloned voice — returns a downloadable WAV URL. No video required.

Batch & Manage

Process multiple videos at once. List, edit, and manage your video library.

Available voices

AI Voice Presets

male1 Default · Fastest

female1 Default · Fastest

male2, male3, female2, female3, female4 — additional styles

Voice Cloning

Provide a voice_sample (local file path or URL to an audio file) and your voice is cloned. Works with narrate_video_full, continue_to_full_video, and generate_tts.

Get started

Create your API key

Connect your AI assistant

Add to .cursor/mcp.json or .windsurf/mcp.json:

Go to Claude.ai → Settings → Integrations → Add MCP Server

Enter the URL and header:

URL:

Header: x-api-key: YOUR_API_KEY

Start a new chat. Local file uploads require a public URL (Google Drive, Dropbox).

Run in your terminal:

Open ⋯ → Manage MCP Servers → View raw config and add:

Add to your MCP config file. Requires Python 3.11+.

Try it

Restart your IDE, then say: "Narrate this video with the male1 voice: <url>"

Available tools

Teach your AI the workflows

Optional but recommended

The MCP tools give your AI assistant capabilities. The NarrateAI Skill teaches it when to use each tool, what to ask you, and the best workflow for your goal — so you get better results without memorizing tool names.

What the skill adds

Picks the right tool for each request automatically
Asks for voice preference (AI preset or clone) before narrating
Handles multi-step flows like translate → re-narrate
Knows batch processing patterns and language requirements

Install by platform

cp -r narrateai-video-workflows .cursor/skills/

1. Zip the narrateai-video-workflows folder
2. Claude.ai → Settings → Capabilities → Skills
3. Upload the zip and toggle it on

cp -r narrateai-video-workflows .claude/skills/

Copy the contents of SKILL.md into your
IDE's AI rules or custom instructions file.

Get the skill on GitHub →

Troubleshooting

API key required Set NARRATEAI_API_KEY in your config or pass api_key to the tool.

Timeout Long videos may exceed the default wait. Use get_job_result(job_id) to poll manually.

429 Rate limit reached. Check your subscription tier and remaining API credits.

Tools not showing Fully restart your IDE (quit and reopen). Check the MCP logs for connection errors.

Ready to start?

Free tier includes 5 minutes of processing. No credit card required.