# 2027.dev

> AI Agent Experience (AX) evaluations for developer tools. We run automated agents against your docs and product, then generate detailed reports with scores, friction points, and recommendations.

## Docs

- [Getting Started](https://docs.2027.dev): Overview of 2027.dev and how it works
- [Quickstart](https://docs.2027.dev/quickstart): Run your first AX evaluation in under 5 minutes
- [MCP Server](https://docs.2027.dev/integrations/mcp): Connect 2027.dev to Claude Code, Cursor, or any MCP client — no install needed, HTTP Streamable transport
- [API Keys](https://docs.2027.dev/integrations/api-keys): Create and manage API keys (starts with evals_)
- [GitHub Integration](https://docs.2027.dev/integrations/github): Auto-run evals on every pull request
- [Understanding Reports](https://docs.2027.dev/reports/overview): What's in an eval report
- [Scoring](https://docs.2027.dev/reports/scoring): How grades (A+ to F) and scores (0-100) are calculated
- [Session Trace](https://docs.2027.dev/reports/session-trace): Replay the agent's evaluation session
- [API Reference](https://docs.2027.dev/api-reference/introduction): MCP server tools and HTTP API

## MCP Server (quick setup)

The 2027.dev MCP server uses HTTP Streamable transport. No npm install needed.

Endpoint: POST https://2027.dev/evals/api/mcp
Auth: Authorization: Bearer evals_YOUR_API_KEY

### Install with one command (Claude Code)

```bash
claude mcp add 2027-evals --transport http https://2027.dev/evals/api/mcp --header "Authorization: Bearer evals_YOUR_API_KEY"
```

Verify: `claude mcp list`

### Or add manually to config file

Claude Code (~/.claude/claude_desktop_config.json) or Cursor (.cursor/mcp.json):
```json
{
  "mcpServers": {
    "2027-evals": {
      "url": "https://2027.dev/evals/api/mcp",
      "headers": {
        "Authorization": "Bearer evals_YOUR_API_KEY"
      }
    }
  }
}
```

### Get your API key

1. Sign in at https://2027.dev/evals
2. Go to Settings → API Keys
3. Click Create Key (starts with evals_)
4. Copy it immediately — it won't be shown again

## Available MCP Tools

- list_reports: List eval reports with filtering and pagination
- get_report: Full report details — scores, dimensions, recommendations
- get_session_trace: Agent conversation replay with tool calls
- get_friction_points: Extract friction points from a report
- compare_reports: Side-by-side comparison of multiple reports
- get_trends: Score history over time
- list_prompts: Available eval prompts for your org
- request_eval: Trigger a new evaluation run
- list_eval_requests: Track eval request status