reelapi
Exposes a run_[ffmpeg](/mcp/servers/integrations/ffmpeg) tool that accepts any ffmpeg command (crop, trim, resize, convert, compress, GIF creation, audio extraction, concatenation) and executes it in the cloud, returning the rendered output file into the project's output directory.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
reelsdk-mcp
Let your coding agent edit video: TikTok-style captions, photo slideshows with voiceover, AI-generated shorts and any ffmpeg command, rendered in the cloud and saved straight into your project. No ffmpeg, whisper or GPU on your machine.
> add TikTok captions to ./clip.mp4
> turn the 8 photos in ./trip into a 30 s slideshow with a voiceover and ./song.mp3 underneath
> make a 45 s short about the Mary Celeste in the dark-documentary style
> crop ./demo.mp4 to 9:16 for ReelsThe agent calls a tool, local files are uploaded, the job runs on ReelSDK, and the finished files land in ./reelsdk-output.
Setup
You need Node 20+ and a ReelSDK API key (rsdk_live_...).
Claude Code
claude mcp add reelsdk --env REELSDK_KEY=rsdk_live_xxx -- npx -y github:bitgate/reelsdk-mcpAdd --scope project to share it with your team through .mcp.json.
Cursor
.cursor/mcp.json in your project (or ~/.cursor/mcp.json for every project):
{
"mcpServers": {
"reelsdk": {
"command": "npx",
"args": ["-y", "github:bitgate/reelsdk-mcp"],
"env": { "REELSDK_KEY": "rsdk_live_xxx" }
}
}
}Claude Desktop
Settings → Developer → Edit Config (claude_desktop_config.json). Desktop has no project folder, so pass absolute paths and pick where results go:
{
"mcpServers": {
"reelsdk": {
"command": "npx",
"args": ["-y", "github:bitgate/reelsdk-mcp"],
"env": { "REELSDK_KEY": "rsdk_live_xxx", "REELSDK_OUTPUT_DIR": "~/Movies/reelsdk" }
}
}
}VS Code
.vscode/mcp.json (asks for the key once and stores it encrypted):
{
"inputs": [{ "type": "promptString", "id": "reelsdk-key", "description": "ReelSDK API key", "password": true }],
"servers": {
"reelsdk": {
"type": "stdio",
"command": "npx",
"args": ["-y", "github:bitgate/reelsdk-mcp"],
"env": { "REELSDK_KEY": "${input:reelsdk-key}" }
}
}
}Windsurf
Settings → Cascade → MCP Servers → View raw config, then add the same mcpServers entry as Cursor.
Related MCP server: Lazynext MCP Server
Tools
Tool | What it does | Typical time |
| Burns animated word-by-word captions into a video (presets | 15-60 s |
| Images to mp4 with transitions, Ken Burns motion, text-to-speech or recorded voiceover, background music ducked under the voice, and captions. | 15-90 s |
| Topic or script in, finished short out: AI script, AI images, voiceover, captions and music, plus a poster frame and a ready-to-post caption. | 1-2 min |
| Any ffmpeg command: crop, trim, resize, convert, compress, GIFs, extract audio, concatenate. | seconds to minutes |
| Keeps waiting on a job that outlived the tool call and saves its files when it's done. | up to |
| Uploads a local file (max 100 MB) and returns a public URL valid for 7 days. Rarely needed: the other tools upload local paths themselves. | upload time |
Every media argument takes a local path (absolute or relative to the project) or a public http(s) URL. Results are never overwritten (clip-captioned-2.mp4), and subtitle files go into reelsdk-output/captions/ so players don't stack them on top of the burned-in captions.
Tools wait up to REELSDK_WAIT_S seconds, sending progress updates if your client shows them. A job that takes longer comes back as a job_id, and the agent calls get_job to pick it up.
Configuration
Variable | Default | |
| Required. Your | |
|
| Where results are saved, on this computer (not a network share). Relative paths resolve against the project root and may not pass through a link; use an absolute path to save elsewhere. |
|
| Seconds a tool waits for its job before handing back a |
|
| API endpoint. https only; plain http only for localhost. |
The project root is CLAUDE_PROJECT_DIR when Claude Code sets it, else the client's first workspace root, else the directory the server was started in. Roots on another machine are skipped.
Limits
Local files up to 100 MB (pass a public URL for bigger inputs, up to 500 MB). Per job: up to 15 minutes of media, 10 minutes of processing, and 20 files or 2 GiB of results to save. A download that sends nothing for 60 s, or runs slower than about 64 KiB/s, stops; the file stays on ReelSDK and its link is in the result.
A local file is only uploaded when it starts like a video, image, audio or subtitle file. Hidden files and folders, meaning names that start with a dot (.env, ~/.ssh), and paths on another machine (\\server\share) are refused.
Results are only downloaded from the API's own /f/ links, and never saved through a link inside the project or the output folder. Node can't open or rename relative to a folder handle, so another program running as you could still swap the folder for a link between the last check and the final rename.
Development
npm ci
npm test # unit + tool tests against a fake API
npm run typecheck
npm run bundle # rebuilds dist/index.js, which is committeddist/ is committed so npx github:bitgate/reelsdk-mcp starts without a build step. See AGENTS.md.
License
MIT
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
AI video editor your agent can drive: plan scenes, edit a multi-track timeline, and render MP4s.
FFmpeg as a service for AI agents: typed video editing tools, async jobs, downloadable outputs.
AI video agent: edit timelines, captions, voice, music, images and ads, then render MP4s.
Agentic AI video editor: edit a real timeline, generate clips, music and voice, then render.
Related MCP Servers
- FlicenseAqualityBmaintenanceEnables AI agents to edit video using text-based proxies, motion graphics via Hyperframes, and advanced FFmpeg rendering, turning any LLM IDE into a professional video editor.4-
- FlicenseNot gradedqualityBmaintenanceEnables AI agents to read, edit, and export real video projects with a full multitrack timeline, effects, captions, and export capabilities.2-
- AlicenseNot gradedqualityCmaintenanceEnables AI coding agents to generate cinematic video and image assets and perform autonomous timeline assembly, multi-track audio editing, chroma-key compositing, and local draft manipulation from conversational prompts.3MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to turn scripts into narrated videos in YouTube 16:9 or Reels 9:16 formats using HTML pages, Remotion compositions, or HyperFrames layouts, with TTS-driven timing, captions, and optional S3 uploads.MIT