Xora

Xora puts FFmpeg behind one production-ready API for developers, automated workflows, and AI agents. Submit raw FFmpeg arguments or choose a recipe, provide media through HTTPS or a presigned object-storage URL, and receive a job ID immediately. Xora runs the work in an isolated cloud worker, exposes explicit progress and terminal states, and returns the result through a signed download URL or your own Cloudflare R2 bucket. The service preserves the flexibility of FFmpeg rather than hiding it behind a narrow template system. It supports normal codec flags, filter graphs, multi-input commands, image and audio outputs, reusable presets, and recipes for common work such as video compression, thumbnails, trimming, concatenation, audio extraction, and web-ready delivery. Probe mode provides enriched ffprobe metadata before processing. Xora is also designed for machine callers. The REST API uses serializable JSON arrays instead of shell strings, failures return structured errors, idempotency keys make resubmission safe, and webhooks report terminal states. An OpenAPI specification, llms.txt reference, and remote MCP server let agents discover and operate the same job system people use from curl or the dashboard. This removes the operational work normally attached to FFmpeg: large binaries, container images, CPU sizing, temporary-disk limits, serverless timeouts, queues, retries, autoscaling, and long-running workers. Teams keep the commands they already understand while Xora supplies the machines and predictable job contract.