Rendobar

About Rendobar

Rendobar is a serverless media processing API that runs jobs such as FFmpeg commands, animated word-level video captioning, static subtitle burning from SRT or VTT files, and styled QR code generation on isolated cloud workers. Jobs are submitted via API call with results returned by webhook, SSE streaming, or polling, and each job runs in an ephemeral, sandboxed container that is destroyed on completion. It offers both CPU and GPU compute options sized to each job.

The product is aimed at developers who need video and media processing capabilities without provisioning, scaling, or maintaining their own infrastructure. It provides a typed JavaScript SDK on npm, a CLI tool called rb, and a webhook-based event system. It also supports AI agent integration through two Model Context Protocol servers: a hosted version at api.rendobar.com/mcp exposing 11 tools over OAuth for clients like claude.ai, and an npm package (@rendobar/mcp) that runs locally for agents in Claude Desktop, Cursor, and Zed to upload local files into jobs.

Billing is per job based on compute time. The free tier includes $5 in one-time signup credits with no card required and stores outputs for 7 days; the Pro tier adds monthly credits, a bonus on top-ups, and 30-day output storage. Credits do not expire. Media is not retained beyond each ephemeral job run, and Rendobar states it does not train models on user media.

Key features

• Run arbitrary FFmpeg commands in the cloud with no servers to provision or scale
• Burn animated word-by-word captions onto video, auto-transcribed and styled
• Burn static subtitles from SRT or VTT files
• Generate styled QR codes
• Compose jobs via timeline (Compose API), webhooks, or SSE streaming instead of polling
• MCP servers (hosted and local npm package) let AI agents like Claude and Cursor call jobs as tools

Frequently asked questions