Hi. I built Instagram to Transcript by myself, mostly in the evenings after my day job. I was writing a blog post recently and wanted to quote a line from a Reel. Instagram draws captions onto the video frame as pixels, so there's nothing underneath to select or copy. I ended up replaying the same ten seconds four or five times, typing while I paused. A public Reel says something out loud, but the words stay locked inside the video unless you sit there with a keyboard. The web tool solved that for one link at a time, and for a while that was the whole project. Then the messages that came in stopped matching what I'd built. A researcher had a spreadsheet of forty Reels pulled from a competitor's account and no interest in pasting them one by one, so I built a page that reads a whole spreadsheet in one pass. A developer wanted the transcript inside a script, as JSON, which became the HTTP API. And when someone described an agent that needed to grab the words from a Reel on its own, the MCP server followed. The API and MCP server both run on credits, a minute of audio to a credit, and a new account starts with ten of them free. A pack is ten dollars for two hundred fifty. I'll be reading every comment today. If you try the API or the MCP server and something breaks while you're wiring it up, tell me here.

Instagram to Transcript

Instagram Reel into text in seconds

Instagram Reel into text in seconds