REST + MCP

Four public-data endpoints behind one API key.

YouTube transcripts for a single video or an entire channel, Google Trends interest and related queries, and Google Play reviews. No browsers to drive, no per-source accounts to keep alive, one response shape to parse.

Subscribe on RapidAPI Use it as an MCP server

Try it before you sign up with the shared demo key wdt_demo_public. It is rate limited and capped at 10 rows per call, and it answers on every tool.

What you get

Every endpoint returns ok, count, seconds and a data array. Nothing silently returns an empty result without a reason.

youtube_transcript

Full transcript text for up to 50 videos in one call, with language, an auto-generated flag, word count, title and channel.

youtube_channel_transcripts

Enumerate a channel, @handle or playlist and get every recent video's metadata and transcript together. Or pass search: your query and go straight from a question to transcripts.

google_trends

Interest over time, interest by region, and top plus rising related queries for up to five keywords.

google_play_reviews

Google Play reviews by app, country and language, including the developer's reply.

REST

Send your key as x-api-key or Authorization: Bearer.

curl "https://web-data-toolkit.vercel.app/api/trends?keyword=moissanite%20ring&geo=US" \
  -H "x-api-key: YOUR_KEY"

curl "https://web-data-toolkit.vercel.app/api/youtube/transcript?url=dQw4w9WgXcQ" \
  -H "x-api-key: YOUR_KEY"
PathRequiredCommon options
/api/youtube/transcripturllang, translateTo, segments
/api/youtube/channelsource — channel, playlist, or search: querymax, lang
/api/trendskeywordgeo, timeframe, byRegion, related
/api/google-play/reviewsappcountry, lang, max, rating

MCP

The same four tools, spoken as Model Context Protocol over HTTP, for Claude, Cursor and any other MCP client.

{
  "mcpServers": {
    "web-data-toolkit": {
      "type": "http",
      "url": "https://web-data-toolkit.vercel.app/mcp",
      "headers": { "x-api-key": "YOUR_KEY" }
    }
  }
}

Built for a context window. An MCP result is spent out of the model's context, so these tools return a reading, not a dump. One Google Trends keyword is 146 rows and about 37,000 characters of JSON; the MCP answer is around 1,200 characters and says where interest sits against the window, the top regions, and the top and rising related queries. Pass raw: true for every field.

Clients that accept only a URL can pass the key in the query string instead: https://web-data-toolkit.vercel.app/mcp?key=YOUR_KEY. The server is stateless, so there is no session to manage.

Which key works here. Use wdt_demo_public to start. A RapidAPI subscription issues a key for RapidAPI's own gateway, and that key does not authenticate against this endpoint, so it will not work for MCP. For a personal key with higher limits, open an issue and say roughly what volume you need.

Speed

Measured through this gateway on 2026-09-14, one call each. Real fetches from the source, not a cache.

CallTimeRows
YouTube transcript, one video3.7s1
Google Play reviews, one app3.8s9
Google Trends, one keyword, US8.6s95
A whole channel, 25 videos with transcripts29.8s26

Those are warm calls. The first call after an idle period pays a cold start, and the slowest single call measured that day was 21 seconds. Set your client timeout to at least 60 seconds.

Honest limits

  1. Public data only. Nothing behind a login, no personal data beyond the public author name a platform already shows on a review.
  2. Transcripts are whatever the platform publishes. Auto-generated captions carry the errors the machine made, and the response tells you which kind you received.
  3. Google Trends numbers are relative interest, the way Google reports them. They are not search volumes.
  4. Review feeds are paginated by the platform, so very large back-catalogues come back in pages rather than all at once.
  5. This is a small independent service. There is no published uptime figure yet because there is not enough history to state one honestly.