Skip to main content
Hive has about 420 provider tools behind it. If a client loaded all of them, the tool list alone would crowd out the conversation, and the agent would have to plan every multi-step job itself. So the root endpoint shows nine tools, each built around one research job.

Nine tools, one per job

When your agent connects to https://mcp.hiveintelligence.xyz/mcp, it sees: The tools page has each one’s inputs, answer shape and price.

What happens inside one call

Each tool runs a fixed set of catalog calls through the same pipeline, in parallel where it can. research_token with the defaults, for example, resolves the name, then fetches market data, top DEX pools, top holders and a GoPlus safety check. Each part comes back as a section with its own receipt. That means:
  • A section that fails is listed under missing. The other sections still answer.
  • A section that does not apply, like DEX liquidity for ETH, says not_applicable and why.
  • If every section fails, the call fails and is refunded.
  • One tool call is one billable call. The catalog calls inside it are not billed separately.
Names are resolved before anything else. When several tokens share a name or ticker, find_token returns ranked candidates, and research_token stops with the candidates instead of reporting on the wrong one.

Reaching the full catalog

The nine tools cover the common jobs. For anything else, connect to the developer profile:
It adds three tools that let an agent reach any catalog tool:
1

Find the right tool

search_tools takes a plain description of the task and returns the tools that match, with each one’s current availability.
2

Read its schema

get_api_endpoint_schema returns the exact parameters, types, and defaults, so the agent builds a valid call instead of guessing argument names.
3

Run it

invoke_api_endpoint executes the call and returns a bounded, typed response with the receipt attached. Each catalog call costs 1 credit.
The CLI reaches the same catalog with hive tools search, hive tools info and hive tools call.

Bounded by default

Each tool trims its sections to keep answers small. Pass detail: "full" for more rows. Catalog tools respect limit, page, per_page, and offset, and Hive caps response size regardless.

Where the data comes from

Hive normalizes many data sources behind one interface. Your agent asks for a price; Hive decides whether that means CoinGecko or an on-chain pool, and tells you which one it used.

Data sources

What each provider covers.