Statey AI – Shared Database for AI Agents Across Every Chat

TL;DR

TL;DR: Statey is a hosted, MCP-native database that acts as shared memory for AI agents — connect it once to Claude, Cursor, or ChatGPT and every session can read and write structured records that persist across chats.

What Is Statey?

Statey is a database your AI uses on your behalf. Unlike a shared notepad or document, Statey is structured: records have schemas, versions, and attributed changes. The core idea is simple — instead of copying context into every prompt, your agents query a shared store that follows them from tool to tool.

The HN launch thread described it well:

“I was a heavy Linear user until I noticed I hadn’t opened the UI in days. I was just asking Claude to pull up the tickets I cared about and draw whatever view I needed in the moment. At some point it clicked that I didn’t need Linear, I needed the database underneath it.”

Statey fills that gap — structured storage that follows the agent, not the human.

Key characteristics:

• MCP-native — connects as a standard MCP server, no custom integration needed
• Hosted — runs as a service, no self-hosted Postgres or schema migrations
• Free to start — $0 tier available, then usage-based
• Multi-agent — works with Claude Desktop, Cursor, ChatGPT, and any other MCP-compatible client

Source and Accuracy Notes

⚠️ This section is MANDATORY. All links must be verified from actual source, not guessed.

• Project page: statey.ai — verified
• Documentation: statey.ai/docs — verified
• HN launch thread: news.ycombinator.com/item?id=48778380 — 4 points, June 2026
• License: proprietary SaaS (no open-source license detected)

Setup Workflow

Step 1: Connect via MCP

Statey registers as an MCP server. In Claude Desktop, add this to your MCP settings:

{
  "mcpServers": {
    "statey": {
      "command": "npx",
      "args": ["-y", "@statey/mcp"]
    }
  }
}

For other MCP clients (Cursor, etc.), use the same package reference or the hosted MCP endpoint documented at statey.ai/docs.

Step 2: Authenticate

# Sign in via browser OAuth
statey auth login

This links your local MCP client to your Statey workspace. The workspace is where records live — think of it as a shared project database.

Step 3: Create Your First Collection

# Via CLI
statey collections create tickets

Or ask your agent directly: “Create a collection called tickets with fields for title, status, priority, and assignee.”

Step 4: Start Using It

From any connected agent session:

Claude: "Track this bug: form validation fails on Safari, high priority, assign to alice"

Statey writes a structured record. Ask a different session later:

Claude: "What high-priority tickets are still open?"

The second session reads from the same shared store — no copy-pasting, no context resurrection.

How It Works Under the Hood

Statey exposes a tool surface over MCP. Every tool maps to a CRUD operation on your workspace collections. The MCP protocol handles the transport; Statey handles the storage.

Key tools available:

| Tool | What it does | |------|-------------| | collections.create | Create a new collection with a JSON schema | | records.put | Write a record into a collection | | records.query | Filter records by field values | | records.update | Patch fields on an existing record | | records.delete | Remove a record |

Schemas are versioned. Every change to a record is attributed to the human or agent that made it, building an audit trail automatically.

Practical Evaluation Checklist

• MCP compatibility — works with Claude Desktop, Cursor, and any MCP v1-compatible client
• No self-hosting — hosted service, no infrastructure to manage
• Structured data — unlike vector stores or document databases, Statey stores typed records with schemas
• Cross-agent sharing — same data visible across all connected agents, not just within one chat session
• Free tier — $0 to start, pricing scales with workspace size and API calls
• Audit trail — every record change is attributed to the actor (human or agent) that made it

Security Notes

• Data is stored in Statey’s hosted environment — review their privacy policy before using with sensitive data
• MCP authentication is per-workspace via OAuth
• No encryption details are currently published on the landing page; enterprise or self-hosted options may be on the roadmap

FAQ

Q: Is this open source? A: No. Statey is a proprietary hosted service. There is no self-hosted or open-source edition mentioned in the current documentation.

Q: How is this different from a shared Google Doc or Notion? A: Statey stores structured, typed records — not free-form text. This means agents can query (“show all open bugs assigned to alice”), not just retrieve a block of context. The data is also machine-readable at the field level, not just human-readable.

Q: Does it work with Claude Code or other CLI agents? A: Any agent that supports the MCP protocol can connect. The npm package @statey/mcp is the standard client.

Q: What happens when the free tier runs out? A: Pricing is usage-based beyond the free tier. Specific per-record or per-query costs are not listed publicly — check statey.ai for the current pricing page.

Conclusion

Statey solves the context-shortfall problem for AI agents in a way that feels natural: instead of managing a database yourself, your agent does it for you. The MCP integration means it slots into existing workflows without custom code.

The core value proposition is cross-session memory without the engineering overhead. For developers who live in Claude Desktop or Cursor and want their agents to “remember” across sessions without manual context management, Statey is worth a look.

Next step: Open statey.ai/docs and connect it to your preferred MCP client in under five minutes.