FeedMCP

Guides

Extract video transcripts from YouTube, TikTok, Instagram, X & Meta Ads for LLMs

FeedMCP gives YouTube, TikTok, Instagram, X, LinkedIn, Reddit and Meta ads one $0.002/call transcript schema with full text and timestamped segments. Missing transcripts or videos over the 2-minute AI-transcription cap are free.

Endpoints used

Endpoint Path Price
YouTube transcript GET /v1/youtube/videos/transcript/get $0.002
TikTok transcript GET /v1/tiktok/videos/transcript/get $0.002
Instagram transcript GET /v1/instagram/posts/transcript/get $0.002
X video transcript GET /v1/x/posts/transcript/get $0.002
Meta ad transcript GET /v1/meta/ads/transcript/get $0.002

Steps

1. Pick the platform transcript endpoint

All 7 endpoints share the exact same response shape, so one parser handles every platform.

2. Request full text and timed segments

Pass the video URL/ID with X-Max-Price: 0.002 to receive data.item.text and data.item.segments.

3. Handle not_found and video_too_long gracefully

Both return charged: false—skip cleanly instead of retrying in a loop.

Complete code

const API = "https://api.feedmcp.com";
const endpoints = {
  youtube: "/v1/youtube/videos/transcript/get",
  tiktok: "/v1/tiktok/videos/transcript/get",
  instagram: "/v1/instagram/posts/transcript/get",
  x: "/v1/x/posts/transcript/get",
  meta: "/v1/meta/ads/transcript/get",
};

export async function getTranscript(platform, id) {
  const url = new URL(API + endpoints[platform]);
  url.searchParams.set("id", id);
  const res = await fetch(url, {
    headers: {
      Authorization: `Bearer ${process.env.FEEDMCP_API_KEY}`,
      "X-Max-Price": "0.002",
    },
  });
  const body = await res.json();
  if (body.code !== "ok") return null; // not_found / video_too_long are never charged
  return body.data.item;
}

Cost estimate

Transcribing 50 competitor videos or Meta ads costs just $0.10 total—ideal for bulk hook analysis in an LLM.

Questions

How much does it cost to run this guide?

Transcribing 50 competitor videos or Meta ads costs just $0.10 total—ideal for bulk hook analysis in an LLM.

Can I do this without writing code?

Yes. Install the FeedMCP MCP server in Claude, Cursor or Cherry Studio and run the same workflow in plain language.