Skip to main content
Fincept MCP has hundreds of tools. Listing all of them would fill your agent’s context before it starts, so the server lists a core set and three discovery tools. Every other tool stays callable by name.

​
What is listed by default

  • Core tools: the most used tool of each area, for example market_get_quote, market_get_candles, data_search_series, stats_ols, stats_arima.
  • Discovery tools: fincept_search_tools, fincept_describe_tool, fincept_call_tool.
  • Engine catalogs: stats_catalog, forecast_catalog, alloc_catalog, pfopt_catalog and quant_catalog list each engine’s modules and tools.
  • fincept_read_result: pages through large results. See results.

​
How the agent finds the rest

1

Search

fincept_search_tools takes plain words and returns matching tools with their signature and what they do.
2

Describe

fincept_describe_tool returns one tool’s full description and JSON input schema.
3

Call

The agent calls the tool by name, exactly like a listed tool. Clients that only call listed tools can use fincept_call_tool:
Agents follow this pattern on their own; the server’s instructions tell them to. Naming the area in your prompt (“using Fincept’s filings tools…”) speeds it up.

​
List whole toolsets

Add toolsets to the endpoint URL to list more tools up front. Use all for everything, or a comma-separated list:
Unknown names are ignored. Platform toolset names are in the toolsets reference. Engine toolsets are named <prefix>_<module>, such as stats_tsa_models, forecast_backtest or alloc_parity, and listed in each engine’s reference: Fincept Stats, Fincept Forecast, Fincept Allocation, Fincept Portfolio Optimizer and Fincept Quant.
all lists several hundred tools. Many clients cap or slow down with that many; prefer a short list of the toolsets you use.

​
Tool annotations

Each tool declares MCP annotations your client can use for approvals: Tools that write to your account (saving a watchlist item, placing a paper order) say so in their description and annotations.