Rendobar Team

Introducing Rendobar: Serverless Media Processing API

Watermark, transcode, caption, and render videos with one API call. Rendobar is a serverless media processing platform built on Cloudflare Workers with credit-based billing and MCP support for AI agents.

Today we are launching Rendobar, a serverless media processing API designed for developers who need to process video at scale without managing infrastructure.

The problem

Video processing is complex. Setting up FFmpeg, managing GPU servers, handling file storage, building job queues, and tracking usage all take significant engineering effort. Most teams end up building and maintaining custom infrastructure that distracts from their core product.

Our approach

Rendobar wraps all of this into a single REST API. You submit a job with a video URL and parameters, and we handle the rest: downloading, processing, uploading, and notifying you when the result is ready.

What you can do today

At launch, Rendobar supports six core job types:

  • Watermark videos with image or text overlays, configurable position, scale, and opacity
  • Transcode and convert between formats with full codec control
  • Generate thumbnails from any video timestamp
  • Extract captions using AI-powered speech-to-text transcription
  • Burn captions into video with customizable styling
  • Run raw FFmpeg commands with safety guardrails

We also support AI-powered compliance workflows including content moderation, face detection, and automatic GDPR redaction.

How it works

The architecture is Cloudflare-first. The API runs as a Cloudflare Worker using Hono, with D1 for the database, R2 for storage, KV for caching, and Queues for status updates. Heavy processing jobs are dispatched to Trigger.dev execution providers.

Every job follows the same flow:

  1. Submit a job via the REST API
  2. Rendobar validates, checks credits, and dispatches to an execution provider
  3. The provider processes the video and uploads the result to R2
  4. You get notified via polling, SSE, or webhooks

Credit-based billing

Pricing is simple: every plan includes a monthly usage balance. Jobs are priced by video duration or a flat rate. Unused balance rolls over (up to 2x your monthly allocation). There is a free tier with $5 of usage per month so you can try everything without a credit card.

MCP-native

Rendobar is built for AI agents from day one. The API includes 15 MCP tools that map 1:1 with the REST endpoints. AI agents can authenticate via OAuth through the Better Auth MCP plugin, then submit jobs, check status, and manage resources autonomously.

What is next

We are actively building more job types, client SDKs, and additional render engine integrations. The roadmap includes Remotion rendering, pipeline composition (chaining jobs), and a visual template editor.

If you are building something that involves video processing, we would love to hear from you.