web_search
Searches the web for current information. Uses a multi-strategy fallback chain to ensure reliability across different hosting environments:
- Firecrawl (primary) — structured search API, most reliable
- DuckDuckGo Lite — POST-based HTML scraper, works from cloud IPs
- DuckDuckGo HTML — legacy fallback
- Bing — HTML scraper fallback
- Wikipedia — final fallback for factual queries
query (required)
Returns up to 5 results, each with title, url, and snippet.
web_extract
Fetches a URL and extracts the readable text content. Strips scripts, styles, navigation, and footer elements. Truncates output to 3,000 characters to keep context manageable.
Parameters: url (required)
read_tweet
Reads a tweet from X/Twitter using the FXTwitter API. Accepts a full URL or numeric tweet ID.
Parameters: tweet (required — URL or ID)
photos, videos, and quote fields when the tweet contains media or is a quote tweet.
find_music
Searches for music on YouTube and returns direct video links. Uses Firecrawl with a music-biased query, falling back to DuckDuckGo.
Parameters: query (required)
get_crypto_price
Fetches live cryptocurrency prices from CoinGecko. Supports common ticker aliases (e.g., eth → ethereum, btc → bitcoin).
Parameters: coin (required)
Built-in aliases:
| Input | CoinGecko ID |
|---|---|
eth, ethereum | ethereum |
btc, bitcoin | bitcoin |
sol, solana | solana |
usdc | usd-coin |
link | chainlink |
uni | uniswap |
pepe | pepe |
doge | dogecoin |