MCP server / Instagram

Instagram MCP server: Reels as text, inside your AI

Paste a Reel or video-post link and your assistant works with the spoken words, no download, no re-upload, no caption export needed.

Connect it to your client

Any MCP client connects the same way: point it at the server URL and send a TranscriptFetch API key as a bearer header (on claude.ai, OAuth replaces the key). Client-by-client walkthroughs: Claude, Cursor, Codex, or the configuration reference.

Server URL
https://transcriptfetch.com/mcp
Any stdio client (mcp-remote bridge)
npx mcp-remote https://transcriptfetch.com/mcp \
  --header "Authorization: Bearer tf_live_YOUR_KEY"

What it can do

get_transcript takes an instagram.com/reel/… or /p/… link and returns timestamped text. That makes the classic Reel workflows one prompt long: repurpose your own Reel into a caption, a post, or a newsletter section; or hand an agent a set of competitor Reels and ask what they are actually saying.

In your assistant, after connecting
You: Turn this Reel into a LinkedIn post in my voice:
     instagram.com/reel/...

Agent: [get_transcript]
       Draft: The thing nobody tells you about...

Why Instagram text is locked up

Instagram never publishes captions as something you can export. What you see on a Reel is rendered for viewing only, and the official Instagram APIs are built for publishing and insights, not for reading back what a video said. So the only reliable route to the words is the audio itself, which is exactly what this server transcribes.

The distinction that trips people up is caption versus captions: the text written under a post is not what this returns. This returns the spoken words from the video. For copying the written caption there is a browser trick, covered in our blog, but an assistant working with the content of a Reel almost always wants the speech.

Limits worth knowing before the first prompt: private accounts are off the table by design, Stories vanish after 24 hours, and a Reel whose message was typed into the frames rather than spoken will transcribe to whatever the audio held, which may be nothing.

Frequently asked questions

Does it return the written caption under the post?

No, it returns the spoken words from the video's audio. The written caption is visible in a browser and copyable there; the speech is what needs transcription, and that is this server's job.

Which Instagram links work?

Public Reels, video posts and IGTV uploads, via instagram.com/reel/… or /p/… links. Private accounts are not accessible, and Stories only exist for 24 hours.

Do Reels with auto-captions transcribe faster?

Instagram exposes no caption file either way, so every fetch is an AI transcription of the audio. Reels are short, so it is seconds, not minutes.

What does it cost?

100 free credits a month, one credit per successful transcript, failures and empty results never billed. Paid plans start at $5 for 500 credits.

Reels in, text out

Sign up, grab a key, paste one config block into your client. Your assistant reads Instagram from then on.

Instagram MCP Server: Reel Transcripts for AI Clients