What is Xora?
Xora is a cloud FFmpeg API for developers and AI agents that need media processing without running their own encode infrastructure. Send raw FFmpeg arguments, a recipe, or a saved preset with an input URL, then poll the job or receive a webhook when the output is ready. Jobs run on isolated workers, support HTTPS inputs including S3 and R2 presigned URLs, and can return signed downloads or write results to Cloudflare R2.
Why Xora works
FFmpeg becomes awkward when it leaves a developer laptop because serverless limits, sandbox constraints, and long running encodes turn one command into a queue, worker, storage, and retry problem. Xora keeps the FFmpeg command but moves the compute behind one job API with explicit state, parallel workers, idempotency keys, and structured errors. It runs the media job on managed workers instead of making the developer operate a media pipeline.
Xora features
- Raw FFmpeg commands. Send FFmpeg arguments as a JSON array, including filter graphs, codec parameters, and multi input jobs, without translating them into a narrower media API.
- Recipes and presets. Use recipes for tasks such as compression, trimming, thumbnail generation, audio extraction, and concatenation, or save a configuration and rerun it by preset ID.
- Job state and webhooks. Jobs expose states and progress that code can inspect, with polling or webhooks available for completion handling. Idempotency keys support safe retries.
- Flexible input and output storage. Read media from HTTPS URLs, including presigned S3 and R2 URLs, upload local files through presigned uploads, use managed storage with signed URLs, or send outputs directly to Cloudflare R2.
- Agent friendly access. The same service is available through REST, OpenAPI, llms.txt, and a remote MCP server with tools for creating and inspecting jobs.
- Credit based plans. The free plan includes 25 monthly credits, a 250 MB input limit, and two concurrent jobs. Build costs $15 per month with 400 credits and a 2 GB limit, while Scale costs $49 per month with 1,500 credits and a 6 GB limit.
Who Xora is for
- Developers adding video or audio processing to an app who do not want to maintain FFmpeg workers, queues, and autoscaling.
- AI agent builders who need media operations exposed through structured JSON, MCP tools, and deterministic job states.
- Developers moving FFmpeg workloads off serverless functions because package, disk, CPU, or execution limits interfere with longer jobs.
- Teams already using FFmpeg commands locally who want to run the same arguments on remote compute without adopting a new media abstraction.
Similar micro SaaS ideas you can build
- Background document conversion API. An API for developers who need heavy file conversions that regularly outgrow serverless limits. It accepts a conversion job, runs the required binary on isolated workers, and reports structured state back to the calling app.
- Agent callable batch processing service. A small infrastructure product that gives AI agents safe access to long running command line workloads through typed arguments, explicit job states, retries, and webhooks instead of direct shell execution.
- Podcast production job runner. A processing API for podcast apps that accepts uploaded recordings and runs repeatable audio tasks such as extraction, conversion, and preset based preparation without requiring the app team to maintain media workers.
