> ## Documentation Index
> Fetch the complete documentation index at: https://docs.core3.io/llms.txt
> Use this file to discover all available pages before exploring further.

# CORE3 MCP

> Connect AI agents to CORE3 risk intelligence through the public core3 server—read-only category scores, project metrics, Proof of Voice, and exchange scores over Streamable HTTP.

The **Model Context Protocol (MCP)** is an open standard that lets AI agents call external tools and retrieve structured data during a conversation. CORE3 exposes a public MCP server named **`core3`** so you can integrate category quality scores, Proof of Voice (PoV), and related risk metrics into agent workflows without building a custom REST integration.

<Info>
  MCP is defined by the open specification at [modelcontextprotocol.io](https://modelcontextprotocol.io). This documentation covers CORE3's server—not the full protocol tutorial.
</Info>

## What CORE3 offers

Through **`core3`**, you get **23 read-only tools** that return CORE3 risk intelligence:

* **Project slug resolution** — map brand names and tickers to canonical CORE3 slugs via `resolve_project_slug`
* **Project category quality scores** — security, financial, operational, reputational, and regulatory (higher = better; not overall PoL)
* **Project metrics** — age, GitHub activity, certifications, red flags, social fraud signals, public registration, and market data
* **Proof of Voice** — expert reviews and structured pros/cons
* **Exchange category quality scores** — security, transparency, and solvency (higher = better)
* **Historical trends** — category score time series for projects and exchanges

Every tool is read-only. No write or mutation operations are exposed.

<Note>
  Overall Probability of Loss (PoL) — where lower = safer — is available through the REST APIs, not as an MCP tool. All MCP score tools return **category quality** scores (`scoreType: "category_quality"`, higher = better).
</Note>

## Server identity

| Property       | Value                          |
| -------------- | ------------------------------ |
| Server name    | `core3`                        |
| Version        | `1.0.0`                        |
| Endpoint       | `https://api.core3.io/v1/mcp`  |
| Transport      | Streamable HTTP (stateless)    |
| Authentication | **None** — no API key required |

## Scope and limitations

<Note>
  MCP exposes a **curated public subset** of CORE3 data. The full REST APIs offer search, ratings pagination, history by timestamp range, exchange dashboards, overall/index PoL, and other endpoints not available as MCP tools. See [Projects Data API](/projects-data-api) and [Exchanges Data API](/exchanges-data-api) for the full programmatic surface.
</Note>

The public tool set may evolve as CORE3 expands its agent integration offering. Rate limits are subject to fair-use policies; contact CORE3 for production-scale REST access.

<Columns cols={3}>
  <Card title="Getting started" icon="rocket" href="/mcp-getting-started" cta="Connect your client">
    Configure Cursor, Claude Desktop, or another MCP client
  </Card>

  <Card title="Tools reference" icon="list" href="/mcp-tools-reference" cta="Browse all tools">
    Full catalog of 23 tools with parameters and REST mappings
  </Card>

  <Card title="Project PoL methodology" icon="flask-vial" href="/project-pol-methodology" cta="Learn scoring">
    Understand how category scores are calculated
  </Card>
</Columns>

<Tip>Need full REST API access, higher rate limits, or commercial terms? Contact [info@core3.io](mailto:info@core3.io)</Tip>

## Next steps

* Follow [Getting started](/mcp-getting-started) to connect your MCP client and run your first tool calls.
* Use [Tools reference](/mcp-tools-reference) when building agent prompts that need specific CORE3 data.
* Request REST API access via [CORE3 for partners](/core3-for-partners) if you need the full programmatic surface or commercial terms.