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

# Cursor

> Integrate Negentro directly into Cursor using the standard stdio MCP transport to provide full access to 30 cognitive memory tools, semantic search, and context assembly.

## MCP Server Configuration

Add the Negentro MCP server to your Cursor configuration file ( .cursor/mcp.json or global Cursor settings):

```json JSON theme={null}
{
  "mcpServers": {
    "negentro": {
      "command": "npx",
      "args": ["-y", "@piyapi/mcp-server"],
      "env": {
        "PIYAPI_API_KEY": "sk_live_your_api_key_here"
      }
    }
  }
}
```

## Available MCP Tools in Cursor

Once connected, Cursor's AI composer and chat can invoke 30 specialized memory tools:

| TOOL NAME | PARAMETERS | CLIENT DESCRIPTION |
| :- | :- | :- |
| store\_memory | content, tags, metadata | Store memory with vector embedding and graph links. |
| get\_memory | id | Retrieve specific memory object by UUID. |
| search\_memories | query, limit, min\_score | Hybrid search (dense vector + BM25 trigram). |
| fuzzy\_search | query, limit | Typo-tolerant trigram search for misspelled input. |
| get\_context | query, max\_tokens | Token-budget-aware context assembly for prompt. |
| ask\_memory | question, temperature | RAG-powered Q\&A with structured citations. |
| pin\_memory | id, is\_pinned | Pin/unpin memory in LLM context window budget. |
| graph\_traverse | start\_id, max\_hops | Perform 1-3 hop graph traversal across linked entities. |

## Agent Skills Framework

Cursor automatically inspects repository-level skills from .agents/skills/data-connectors/SKILL.md to discover connected systems and credentials.

```bash Terminal theme={null}
# Cursor Skill discovery path
.agents/skills/data-connectors/SKILL.md
```


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.