Skip to main content
The Tavily connector lets your workflows search the web, extract page content, map and crawl sites, all through Tavily’s hosted MCP server. Useful for grounding AI agents with fresh information that isn’t in their training data.
Spojit does not curate the tool list; Tavily does. For the live set of tools and their parameters, see the Tavily MCP documentation.

​
Connection setup

1

Get your API key

Sign in to app.tavily.com and copy your API key from the dashboard. Tavily keys start with tvly-.Tavily has a free tier with a monthly request quota; check current limits and pricing on tavily.com/pricing.
2

Add the connection in Spojit

Go to Connections in Spojit, click Add Connection, select Tavily, and paste your key into the API Key field.

​
When to use Tavily

  • Grounding AI agents: search the web from inside a workflow to give the model fresh facts.
  • Content extraction: pull the readable body out of a URL without writing scraping logic.
  • Site mapping / crawling: discover URLs across a domain.
If you only need occasional grounded answers in chat, the built-in chat web-search (powered by Gemini grounding) covers that without a Tavily connection. Reach for the Tavily connector when you want web access from inside a workflow, not just chat.

​
Caveats

  • API key only. There is no OAuth flow; the key fully represents your account. Treat it like a password and rotate it on Tavily’s dashboard if exposed.
  • Usage is metered by Tavily. Each search/extract/crawl call consumes from your monthly quota. Monitor usage in the Tavily dashboard; once the quota is exhausted, calls fail until the next billing cycle (or until you upgrade).
  • MCP endpoint is hosted by Tavily. Availability and the exact tool surface are determined by Tavily, so outages and breaking changes are upstream.

​
Tools

For the current list of tools and their parameters, see Tavily’s MCP documentation.