# AgentLair

> AgentLair is a freemium MCP server for communication. Gives agents a persistent email address, encrypted vault, audit trail and trust score. It can run locally on your machine or as a hosted remote server.

- **Kind:** MCP server
- **Pricing:** Freemium
- **Platforms:** API, CLI
- **Website:** https://agentlair.dev
- **Repository:** https://github.com/piiiico/agentlair
- **MCP Registry name:** io.github.piiiico/agentlair
- **Categories:** communication, security
- **Listing updated:** 2026-11-14

## About

The idea is that an agent's identity should outlast its session. Registering takes one call to /v1/auth/agent-register, which returns an API key, an account ID and an address such as my-agent@agentlair.dev on the free tier, limited to 10 emails and 100 requests a day. Email is sent and received over REST with no SMTP, OAuth or CAPTCHA, with drafts, threading and webhooks.

Credentials go in a vault where the agent encrypts with AES-GCM on the client side and the server stores only ciphertext. Every action can be logged with an Ed25519 signature into a tamper-evident chain, and a behavioural trust score from 0 to 100 is derived from observed consistency, restraint and transparency rather than what an agent says about itself. A public demo endpoint shows healthy, suspicious and new-agent scores without signup.

The site also describes an Agent Authentication Token, signed with EdDSA and verifiable through JWKS, and isolated pods for multi-agent or multi-tenant use. The MCP package @agentlair/mcp runs over stdio with an AGENTLAIR_API_KEY, or connects to a remote streamable HTTP endpoint. Python and TypeScript SDKs are also published.

## Features

- **Agent email address:** Send and receive at an agentlair.dev address over REST.
- **Zero-knowledge vault:** Client-side AES-GCM encryption; the server stores only ciphertext.
- **Signed audit trail:** Ed25519-signed, tamper-evident action records.
- **Behavioural trust score:** A 0-100 score built from observed actions.
- **Agent tokens:** EdDSA-signed tokens verifiable through JWKS.
- **Pods:** Namespace isolation per client or agent group.
- **SDKs and MCP:** Python and TypeScript SDKs plus an npm MCP package.

## Good for

- Give a research agent its own inbox
- Keep an agent's credentials out of environment variables
- Record signed evidence of what an agent did
- Check an agent's trust score before accepting its request
- Separate several clients' agents into isolated pods

## Install

### Claude Code (Terminal)

```sh
claude mcp add --transport http agentlair https://agentlair-mcp-server.amdal-dev.workers.dev/mcp
```

### Claude Desktop (Settings → Connectors → Add custom connector)

```sh
https://agentlair-mcp-server.amdal-dev.workers.dev/mcp
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "agentlair": {
      "url": "https://agentlair-mcp-server.amdal-dev.workers.dev/mcp"
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "agentlair": {
      "type": "http",
      "url": "https://agentlair-mcp-server.amdal-dev.workers.dev/mcp"
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "agentlair": {
      "serverUrl": "https://agentlair-mcp-server.amdal-dev.workers.dev/mcp"
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "agentlair": {
      "type": "http",
      "url": "https://agentlair-mcp-server.amdal-dev.workers.dev/mcp"
    }
  }
}
```

## FAQ

### How do I register an agent?

POST a name to https://agentlair.dev/v1/auth/agent-register; the response includes the API key and email address.

### What are the free limits?

The README shows 10 emails a day and 100 requests a day on the free tier.

### Who can read the vault?

The README says encryption happens client-side and the server stores ciphertext only.

### Can I try the trust score without signing up?

Yes, GET /v1/demo returns sample healthy, suspicious and new-agent responses, rate limited to 10 requests a minute per IP.

### Which transports does the MCP server support?

stdio through the @agentlair/mcp npm package and streamable HTTP.

### Is AgentLair free?

AgentLair has a free plan, with paid plans for more usage or features.

### How do I add AgentLair to Claude Code?

Run this command in your terminal: claude mcp add --transport http agentlair https://agentlair-mcp-server.amdal-dev.workers.dev/mcp

### What are the best alternatives to AgentLair?

The closest alternatives to AgentLair are REA, Agent Roast, Midplane, Leash and Tinplate Edge AI Safety, Security & EU Compliance.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Agent Roast](https://ai-powered.com/mcp/agent-roast): Agent Roast is an open-source MCP server for developer tools.
- [Midplane](https://ai-powered.com/mcp/midplane): Midplane is an open-source MCP server for databases.
- [Leash](https://ai-powered.com/mcp/leash): Leash is a free MCP server for security.
- [Tinplate Edge AI Safety, Security & EU Compliance](https://ai-powered.com/mcp/tinplate-edge-ai-safety-security-and-eu-compliance): Tinplate Edge AI Safety, Security & EU Compliance is a freemium MCP server for security.
- [SecureCloud](https://ai-powered.com/mcp/securecloud): SecureCloud is a free MCP server for security.
- [BoardAgent](https://ai-powered.com/mcp/boardagent): BoardAgent is a free MCP server for community directories.
- [Callout](https://ai-powered.com/mcp/callout): Callout is a freemium MCP server for communication.

---
Source: https://ai-powered.com/mcp/agentlair
