caption.sh is an AI captions API for developers. One POST turns a raw video into an MP4 with word-timed, burned-in captions: karaoke-style active-word highlighting, 18 bundled fonts, outline or backdrop styles, and sizing that adapts to any aspect ratio. Async jobs with webhooks, prepaid per-minute billing, and a hosted MCP server so agents can caption video as a native tool. Built for clipping pipelines, UGC tools, and anyone shipping short-form video at volume.
We kept building the same captioning pipeline for every video project: transcription, word timing, styling, rendering. It should have been an API call the whole time, so we built it as one.