Quick Start

RushDB is a graph database built for agents and structured data. Pick the path that fits how you work:

Path	Best for	Time
Web AI connector	ChatGPT or Claude.ai, no install	~1 min
Local MCP	Claude Desktop, Cursor, VS Code	~3 min
Skills	Any MCP-capable agent in VS Code	~2 min
SDK / REST	Custom apps and pipelines	~5 min

Need an API key first? → Get API Key

---

Web AI connector

Connect to ChatGPT or Claude.ai using the hosted MCP endpoint. No local install, no API key needed for web clients — just OAuth.

ChatGPT:

1. Open Settings → Connectors → Add connector
2. Enter URL: https://mcp.rushdb.com/mcp
3. Sign in with your RushDB account

Claude.ai:

1. Open Settings → Integrations → Add integration
2. Enter URL: https://mcp.rushdb.com/mcp
3. Authorize with your RushDB account

Then verify the connection:

"Call getSchemaMarkdown and show me what's in my RushDB project."

→ Full MCP client list and options

---

Local MCP

Works with Claude Desktop, Cursor, VS Code, or any stdio-based MCP client. Requires a RushDB API key.

Claude Desktop — edit ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "rushdb": {
      "command": "npx",
      "args": ["-y", "@rushdb/mcp-server"],
      "env": { "RUSHDB_API_KEY": "your-api-key-here" }
    }
  }
}

Cursor — add to .cursor/mcp.json:

{
  "mcpServers": {
    "rushdb": {
      "command": "npx",
      "args": ["-y", "@rushdb/mcp-server"],
      "env": { "RUSHDB_API_KEY": "your-api-key-here" }
    }
  }
}

VS Code — add to .vscode/mcp.json:

{
  "servers": {
    "rushdb": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@rushdb/mcp-server"],
      "env": { "RUSHDB_API_KEY": "your-api-key-here" }
    }
  }
}

After connecting, bootstrap agent memory in one prompt:

"Set up RushDB as my persistent memory layer."

The agent calls getSchemaMarkdown, creates a SESSION record, and confirms recall is working. Full bootstrap prompt and memory model: rushdb.com/agent-setup.

Verify the connection

Ask: "Call getSchemaMarkdown and show me what labels exist in my project." A working connection returns the schema; an empty project returns an empty schema — both are correct responses.

---

Skills

Install the RushDB skills to give your agent structured domain knowledge — so it knows how to query, model data, and manage agent memory without you explaining it every session.

Compatible with Claude, GitHub Copilot, Cursor, Windsurf, and any Agent Skills-compatible client.

Skill	What it enables
rushdb-agent-memory	Store sessions, decisions, and context; recall by meaning
rushdb-query-builder	Build findRecords filters, aggregations, and semantic searches
rushdb-data-modeling	Design labels, properties, relationships, and nested schemas
rushdb-faceted-search	Build faceted filter UIs from property metadata
rushdb-domain-template	Design a schema for any domain through guided conversation

Install:

npx skills add rush-db/rushdb --path packages/skills

Or via npm:

npm install @rushdb/skills

Source and docs: github.com/rush-db/rushdb — packages/skills

note

Skills work best when the MCP server is also connected. The skills tell the agent how to use RushDB; the MCP server gives it the tools to do so.

---

SDK

For custom apps, data pipelines, or direct programmatic control.

Python

from rushdb import RushDB

db = RushDB('RUSHDB_API_KEY')

# Push data
db.records.create_many(
    label='MEMORY',
    data=[{'content': 'RushDB stores structured memory for AI agents.'}]
)

# Query by meaning
results = db.ai.search({
    'propertyName': 'content',
    'query': 'how agents remember things',
    'labels': ['MEMORY']
}).data

TSTypeScript

import RushDB from '@rushdb/javascript-sdk';

const db = new RushDB('RUSHDB_API_KEY');

// Push data
await db.records.importJson({
label: 'MEMORY',
data: [{ content: 'RushDB stores structured memory for AI agents.' }]
});

// Query by meaning
const { data: results } = await db.ai.search({
propertyName: 'content',
query: 'how agents remember things',
labels: ['MEMORY']
});

shell

curl -X POST "https://api.rushdb.com/api/v1/records/import/json" \
  -H "Authorization: Bearer $RUSHDB_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"label": "MEMORY", "data": [{"content": "RushDB stores structured memory for AI agents."}]}'

→ TypeScript SDK · Python SDK · REST API

---

What to build

Use case	Guide
Agent memory — sessions, decisions, context	Episodic Memory for Multi-Step Agents
GraphRAG — graph + vector search	GraphRAG Tutorial
Semantic search over your data	Semantic Search in 5 Minutes
Hybrid filter + semantic search	Hybrid Retrieval
Explore an unknown dataset	Discovery Queries
Safe agent query planning	Agent-Safe Query Planning

---

Concepts to know

Three ideas explain how RushDB works:

• Records — a node with a label and any JSON properties. No schema required upfront.
• Labels — the record's type (SESSION, DECISION, ARTICLE). You define them by pushing data.
• Relationships — named, directed edges between records. Auto-created from nested JSON.

→ Data model overview · Relationships · Search