Skip to main content
This section covers the AI features in Birdeye Peak. There are two parts:
  • Peak AI — a conversational analyst built into the Peak terminal. Available now to all beta users.
  • For developers — MCP server and agent recipes for integrating Peak data into external AI workflows. Currently in development. Each is documented in its own sub-section below.

​
Peak AI

Peak AI (PAI) is a chat interface inside the Peak terminal that lets you ask crypto-related questions in natural language and get structured live data alongside a written analysis. It draws on Peak’s full dataset and is available to every beta user.

Peak AI Overview

What Peak AI is, how it works, and how to access it.

Using Peak AI

Opening Peak AI, asking questions, and managing credits.

What Peak AI can do

Example questions and use cases Peak AI handles well.

​
For developers

For teams building agents or AI workflows that need programmatic access to Peak data, two integration paths are in development:
  • MCP server — a Model Context Protocol server that will expose Peak Commands as tools, so AI agents can query the same datasets the terminal uses.
  • Agent recipes — reference patterns for common analytical workflows (research agents, monitoring agents, alerting agents) built on top of Peak data.

Developer Integration

Learn more about the planned MCP server and agent recipes, and register interest for early access.

​
How AI tools can read these docs

Separately from the in-product AI features and the planned developer tools, this documentation site is already structured for AI consumption.
FilePurpose
/llms.txtA curated index of all documentation pages, with descriptions. Used by AI tools to discover content.
/llms-full.txtThe full content of every page, concatenated. Useful for tools that want to ingest the entire docs in a single fetch.
/{any-page}.mdAppend .md to any documentation URL to get the clean Markdown version of that page.
The Mintlify platform also hosts an MCP server for this documentation site, separate from the planned Peak product MCP server. The documentation MCP server exposes the docs themselves (concepts, commands, guides) so developers using Claude, Cursor, Windsurf, or other MCP-compatible tools can query these docs from inside their AI workflows.

Data API

The programmatic data layer that developer integrations will be built on.

Object + Command Architecture

The data model that structures both human and AI access to Peak.