# How to get YouTube transcripts into your AI agent

By Alan Lui. Published 2026-09-13; last updated 2026-09-14. Page: https://apifare.com/blog/youtube-transcripts-for-ai-agents

Connect your agent to apifare and ask it to run the Youtube Transcript Scraper from the catalog: give it a video URL and a language, and the transcript comes back to your agent. It's ready to run on your apifare key, so you don't need a YouTube account or API key of your own.

## Copy this line to your agent

```text
Use apifare to run the "Youtube Transcript Scraper" tool with {"videoUrl": "<YouTube video URL>", "targetLanguage": "en"} and summarize the transcript in five bullet points.
```

## Why transcripts, not the video?

Your agent can't watch a video, but it can read one. A transcript turns a talk, tutorial or interview into text it can summarize, quote, compare and search. Video and channel lists let it find the right videos first.

## What does the transcript tool take?

Two fields, both required by its published input schema: `videoUrl` (the video's URL) and `targetLanguage` (for example `"en"`). The full list is on the [tool's page](https://apifare.com/tools/pintostudio/youtube-transcript-scraper).

To find videos first, the [YouTube Scraper](https://apifare.com/tools/streamers/youtube-scraper) takes `searchQueries`, `maxResults` and direct `startUrls`, with an option for transcripts and subtitles.

## Which YouTube tools are on the key?

Transcripts, video search, whole channels, comments and Shorts: the table below maps each job to its tool. The [YouTube API guide](https://apifare.com/apis/youtube) lists the tools most used.

## How is a run priced?

A catalog run reserves a small amount up front, then settles to the run's true cost, never above a fixed ceiling; failed runs are refunded. Each tool's page shows the numbers, and the listed number is what your balance is debited.

## Do I need a YouTube account or API key?

No. Ready-to-run catalog tools run on apifare's own supply credential, so your apifare key is all your agent needs. Use the data within YouTube's terms and applicable law; see [our terms](https://apifare.com/terms).

## For agents: the exact calls

1. Find it: `search_apis` in `apifare-mcp`, or `GET https://apifare.com/v1/search?q=Youtube%20Transcript%20Scraper` with `Authorization: Bearer $APIFARE_TOKEN`. Take the result named "Youtube Transcript Scraper": it carries a `proxy_route` and the tool's `input_schema`.
2. Run it: `proxy_call` with that `proxy_route` and `{"videoUrl": "…", "targetLanguage": "en"}` (apifare-mcp over stdio), or POST the `proxy_route` over HTTP with that JSON as the body and the same header.
3. On HTTP 402 the balance is empty: relay the body's `human_readable` line to your human, wait for the top-up, then retry the same call. Every other error and its fix: [the error catalog](https://apifare.com/errors).

## For humans: set it up

1. **Connect your agent.** Paste `set up https://apifare.com/SKILL.md` into it: it gets a key ($5 of calls free), writes its own config and makes a first call. Or get a key at [/start](https://apifare.com/start?origin=blog) and follow your client's [setup guide](https://apifare.com/integrations).
2. **Ask.** `Get the transcript of this YouTube video: <url>`, or `Find recent YouTube videos about <topic> and list their titles and channels`.

## Which tool for which job

| Job | Tool | Runs on |
| --- | --- | --- |
| The transcript of one video | [Youtube Transcript Scraper](https://apifare.com/tools/pintostudio/youtube-transcript-scraper) | ready to run |
| Find videos by search term, with optional subtitles | [YouTube Scraper](https://apifare.com/tools/streamers/youtube-scraper) | ready to run |
| The videos on a channel | [Fast YouTube Channel Scraper](https://apifare.com/tools/streamers/youtube-channel-scraper) | ready to run |
| The comments on a video | [YouTube Comments Scraper](https://apifare.com/tools/streamers/youtube-comments-scraper) | ready to run |
| The Shorts on a channel | [YouTube Shorts Scraper](https://apifare.com/tools/streamers/youtube-shorts-scraper) | ready to run |

More: [every video tool in the catalog](https://apifare.com/tools/c/videos)

## When this is the wrong tool

- **You need your own channel's analytics** (watch time, revenue). Those live in YouTube Studio, not in public data.
- **The recording isn't on YouTube.** These tools read YouTube URLs.
- **You want the text of an ordinary web page.** Use [scrape](https://apifare.com/capabilities/scrape) (see [reading any web page](https://apifare.com/blog/let-your-agent-read-any-web-page)).

## FAQ

### Do I need a YouTube API key?

No. Ready-to-run catalog tools run on apifare’s own supply credential; your apifare key is all you need.

### How is a run priced?

A run reserves a small amount, then settles to the run’s true cost, never above a fixed ceiling. Each tool’s page shows the numbers, and failed runs are refunded.

### Can it get transcripts for a whole channel?

List the channel’s videos with the channel tool, then run the transcript tool on each one. Per-agent caps keep a long list within the budget you set.

### Which agents can run catalog tools?

Agents using apifare-mcp over stdio run them with proxy_call; any agent can POST the tool’s proxy_route over HTTP. See the [setup guides](https://apifare.com/integrations).

### Can I use the transcripts in my product?

Use the data within YouTube’s terms and applicable law; see [our terms](https://apifare.com/terms).

## Give this to your agent

`set up https://apifare.com/SKILL.md`: it takes it from there (key, config, first call). $5 of calls free.
No agent open right now? Get a key at https://apifare.com/start?origin=blog-youtube, then give it to your agent (setup guides: https://apifare.com/integrations).