transcriptfetch
Get API access

Turn the spoken web into structured data.

Extract videos and podcasts into timestamped text and metadata, or list entire channels and playlists. Captions when available, transcription when needed. One consistent response across every supported source.

Get an API keyTry in ChatGPT or Claude
YouTubeTikTokInstagramSpotify

One-paste setup

Start fetching in seconds

No transcript API hype. Just results.
1,700,000+

Successful transcripts fetched and counting

4.0s
Average response
Last 30 days
99.77%
API uptime
Independently monitored
30 days agoToday

The fastest path to the best transcript available.

TranscriptFetch checks the cache, reads published captions, and transcribes the audio only when needed. Every source returns the same clean structure for your product, research, or AI workflow.

Use published captions for the fastest valid result

Transcribe the audio when captions are missing

Return timestamps and metadata in one consistent schema

Charge nothing for failed transcript requests

YouTubeTikTokInstagramPodcastsDirect videoDirect audioChannelsPlaylistsREST APIMCP

Why spoken

The best sources don't write. They talk.

The founder explains it on a podcast. The engineer demos it on YouTube. The customer reviews it on TikTok. What people actually said increasingly exists only as video, and text-first AI pipelines are blind to all of it.

Public source
YouTube video

Why startups succeed

Bill Gross presenting The single biggest reason why start-ups succeed at TED

Primary source, in their own words

youtube.com/watch?v=bNpx7gpSqbY
6:41 videoCaptions or audio
One request
POST/v1/transcripts/video200 · 1.24s
{"ok": true,"data": {"kind": "transcript","platform": "youtube","video_id": "bNpx7gpSqbY","duration": 401,"segments": [{ "start": 41.0, "text": "I studied hundreds of companies" },{ "start": 46.2, "text": "Five factors shaped their outcomes" },{ "start": 52.3, "text": "One factor stood above the rest" },… 43 more]},"usage": { "credits_spent": 1, "balance": 4812 }}
46 segments · 14.2 KB1 credit

Built for work that starts with a link

Monitor channels, build research datasets, and give agents source-grounded access to what people actually said.

Continuous monitoring

Monitor a market, not one video

Follow a channel or playlist, capture each new upload, and keep a searchable record of the claims, people, and timestamps inside it.

TranscriptFetch watches a set of channels, detects a new founder interview, retrieves its captions, normalizes the transcript, and adds the new source to the archive.

Source set
Channel or playlist
One URL defines what to watch.
New material
Videos and episodes
Poll from the latest known item.
Structured record
Text, time, metadata
One response shape for every source.
Destination
Agent, search, report
Use the result wherever the work happens.

Build research datasets

Normalize interviews, reviews, demos, and podcasts into one timestamped schema ready for search, analysis, or RAG.

Interviews, reviews, demonstrations, and podcasts arrive one by one, then become one normalized research dataset.

Give agents primary sources

Claude, ChatGPT, Cursor, and internal agents can retrieve the original spoken source themselves through MCP.

An agent asks what a founder said about pricing. TranscriptFetch fetches the founder interview, returns exact quotes with timestamps, and attaches the primary source to the answer.

What you can pull with one key

Three capabilities behind a single API key and one response shape: the spoken word, the video and playlist listings around it, and an agent-ready way to reach both.

01
Transcripts

Timestamped video transcripts

Any supported video URL comes back as per-segment text with start times, platform captions when they exist, AI transcription when they don't.

POST /api/v2/transcripts/video
{ "url": "youtube.com/watch?v=dQw4…" }
200captions · en · 412 segments
00:12Most of our corpus was video,
00:19which meant most of it was not
00:26searchable, or citable, at all.
02
Channels & playlists

List a channel or playlist

Resolve a YouTube channel or playlist into a paginated list of videos, newest first. Poll with since_video_id and a page with nothing new costs no credits.

POST /api/v2/transcripts/channel
{ "url": "youtube.com/@channel", "limit": 50 }
20050 of 1,284 videos · newest first
Aug 19How we cut our render time in half
Aug 14A field guide to flaky tests
Aug 09The interview that changed my mind
03
Podcasts

Podcast episodes from a link

Paste a Spotify, Apple Podcasts, or RSS episode link. We resolve it to the publisher's own audio and transcribe it, so the show and episode come back named.

POST /api/v2/transcripts/video
{ "url": "open.spotify.com/episode/4rOoJ6…" }
200audio · en · 1,847 segments
showThe Long Game
episodeBuilding in public, five years on
audio58:12 · publisher CDN
04
Agents

Reachable by your agents

A hosted MCP server plus a prebuilt n8n node, so Claude, ChatGPT, Cursor or VS Code can fetch sources themselves with no glue code.

MCP https://transcriptfetch.com/mcp
{ "method": "tools/list" }
200connected · 5 tools
get_transcriptone video, full text
list_channel_videosa channel, paginated
get_creditsbalance, free to call

Fast when captions exist. Reliable when they don't.

Caption tracks take the fast path. When they are missing or a platform changes, our fallbacks keep the same request moving.

99.77%
API uptime
96%
Caption coverage
$0
Per failed call
12 mo
Deprecation notice

Only 54% of short-form videos carry a usable caption track. When captions are missing, TranscriptFetch transcribes the audio instead, so it returns timestamped text for 96% of the same videos.

TikTok41% → 96%
Instagram38% → 94%
YouTube82% → 99%
URLs where platform captions existURLs where TranscriptFetch returns text1.2M public video URLs, last 90 days
With a scraper or OSS library

One fetch path, so a markup change breaks it quietly and someone on your team owns the fix. No captions means an empty result. You rent and rotate the proxies, you eat the cost of failed calls, and every new platform is another integration to maintain.

With TranscriptFetch

Three fallbacks run behind one request: cache, platform captions through rotating infrastructure, then AI transcription. Failures cost nothing and retry on our side. Five platforms come back in one response shape, and platform changes are ours to chase.

Live statusIndependently monitored

Ready to fetch at scale?

100 free credits every month, no card, no sales call. Failed fetches are never billed.

Get an API key →

Every platform, inside your AI

Connect the MCP server once and Claude, ChatGPT, Cursor or Codex fetch these sources themselves. Each page has the setup for your client.

YouTube

Transcripts, video search, and whole channels or playlists, as tools

YouTube MCP server →

TikTok

Paste a link and your assistant reads it, captioned or not

TikTok MCP server →

Instagram

Reels and video posts, transcribed from the audio

Instagram MCP server →

Podcasts

Spotify, Apple and RSS episodes, through the REST API rather than MCP

Spotify, podcasts and MCP →

Guides & product updates

Guide · Aug 18, 2026

How to copy or download Instagram captions (every method)

Need to copy or download Instagram captions? Here are every practical method, what works on mobile and desktop, and when to automate it.

Read guide →
Comparison · Jul 23, 2026

The Best YouTube Transcript APIs in 2026 (Honest Comparison)

Every way to get YouTube transcripts programmatically in 2026, compared honestly: the Python library that gets IP-blocked in the cloud, hosted APIs, free tools with no API, and DIY yt-dlp + Whisper - with real current pricing.

Read guide →
Guide · Aug 19, 2026

Podcast transcript examples and format: one episode, six ways

The same sixty seconds of a two-person episode shown as raw machine output, clean read, full verbatim, timestamped, SRT, WebVTT and JSON, then the four formatting decisions that separate a transcript people read from one they scroll past.

Read guide →

Everything you need to know, upfront

One credit is one successful response: a video transcript, or one listing call (search, channel, or playlist). Requests that fail or return nothing are always free. Every account gets 100 free credits each month, no card required.

Questions? Come find us

A real person answers. No ticket queue.

GitHub

The SDKs, examples and issue tracker. Report a bug, request a platform, or read the code before you depend on it.

Open GitHub →

Contact

Enterprise pricing, integration questions, or anything that needs a person rather than a doc page.

Get in touch →

Put the spoken web to work

100 free credits a month. No card required. Failed fetches are never billed.

Get an API keyRead the docs