Connect Codex to Tavily
Give your Codex agent access to Tavily in Melso — 5 tools it can call while it works through a task.
Set it up in Melso
- 1In Melso, open Settings → Agent Apps and hit Connect on Tavily.
- 2Finish the API_KEY sign-in in the window that opens. Melso stores the connection, not your credentials.
- 3Open your Codex agent, go to its Agent Apps tab, and allow Tavily.
- 4Give the agent a task. It calls Tavily itself when the work needs it.
What Tavily gives your agent
The tools, permissions, and events Tavily exposes — the same ones your agent can use once the app is authorized.
What you sign in with
Tavily connects with API_KEY.
Permissions
Tavily does not break access down into granular permissions.
Tools Tavily exposes
Tavily crawl
Tool to perform intelligent graph-based website crawling with parallel path exploration and content extraction. Use when you need to traverse and extract content from multiple pages of a website following specific patterns or instructions. Supports depth/breadth controls, domain filtering, and natural language instructions for guided crawling.
Tavily extract
Tool to extract and parse web page content from specified URLs using Tavily's extract endpoint. Use when you need to retrieve clean, structured content from web pages with optional image extraction and content reranking based on query relevance.
Get Tavily usage
Tool to retrieve API key and account usage statistics from Tavily. Use when you need to check credit consumption, limits, and per-endpoint usage for search, extract, crawl, map, and research operations.
Tavily map website
Tool to map a website and discover its pages. Use when you need to scan a website and get a structured list of URLs/pages it contains without extracting full content.
Tavily search
Use this to perform a web search via the Tavily API; offers controls for search depth, content types, result count, and domain filtering. Requires an active Tavily connection (401 = auth failure). Rate limit: ~2 req/s; apply exponential backoff on HTTP 429. Results are nested under response_data.results (not a flat list). Subject to HTTP 429 on rapid bursts.