read_tweet
Reads a tweet/post from X (Twitter) given a URL or numeric tweet ID. Uses the FXTwitter API.
Accepted formats
All of these work as thetweet parameter:
1800000000000000000— numeric IDhttps://x.com/user/status/1800000000000000000— x.com URLhttps://twitter.com/user/status/1800000000000000000— twitter.com URL
Response fields
| Field | Type | Description |
|---|---|---|
id | string | Tweet ID |
text | string | Full tweet text |
author | string | Display name |
handle | string | @username |
date | string | ISO timestamp |
likes | number | Like count |
retweets | number | Retweet count |
replies | number | Reply count |
views | number | View count (may be null) |
url | string | Canonical tweet URL |
photos | string[] | Photo URLs (if present) |
videos | string[] | Video URLs (if present) |
quote | object | Quoted tweet (if present): author, handle, text |
Auto-reading from search results
Whenweb_search returns results containing Twitter/X URLs, the gateway automatically calls read_tweet on up to 3 matching URLs without requiring an additional model round-trip.
find_music
Searches for music and returns YouTube links. The query is augmented with music youtube keywords to bias results toward music content.
Search strategy
- Firecrawl — searches with the augmented query, filters results for YouTube video IDs
- DuckDuckGo Lite — falls back to
{query} site:youtube.comif Firecrawl returns no YouTube results
Response
Returns up to 5 tracks, each with:| Field | Type | Description |
|---|---|---|
title | string | Video title (YouTube suffix stripped) |
url | string | Full YouTube watch URL |
video_id | string | YouTube video ID |
snippet | string | Description (from Firecrawl) or empty |