# Ai Architect MCP Spec

> Ai Architect MCP Spec is an open-source MCP server for developer tools. Deterministic checks that catch invented functions and untraceable claims in specs. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI, self-hosted
- **Website:** https://ai-architect.tools
- **Repository:** https://github.com/cdeust/ai-architect-mcp-spec
- **MCP Registry name:** io.github.cdeust/ai-architect-mcp-spec
- **Categories:** developer tools
- **Listing updated:** 2026-11-14

## About

Specs written by an agent often name functions that do not exist, quote latency nobody measured or list acceptance criteria that test nothing. This server runs structural checks without a model in the loop: every acceptance criterion must trace to a requirement, no performance claim passes without a measurement, and given the symbols of your code graph it reports inline-code symbols in the spec that none of them resolve.

The registry description counts 17 tools covering PRD verification and host-driven generation with multi-judge consensus. The README says it works in Codex, Gemini CLI, Claude Code and CI, runs on your machine and is written in TypeScript. It is one of three standalone servers from the same author, alongside a memory server and a code-graph server.

It installs from an mcpb bundle over stdio. The wider site notes the spec project is still in development.

## Features

- **Requirement traceability:** Checks acceptance criteria against requirements.
- **Measurement rule:** Rejects performance claims without a measurement.
- **Symbol resolution:** Flags spec symbols missing from your code graph.
- **Deterministic verifier:** No model judges the output.
- **Multi-judge consensus:** Generation can be checked by several judges.
- **CI friendly:** Works in Codex, Gemini CLI, Claude Code and CI.

## Good for

- Verify an AI-drafted PRD before it feeds code generation
- Catch references to functions that do not exist
- Gate spec quality in CI

## Install

### Claude Code (Terminal)

```sh
claude mcp add ai-architect-mcp-spec -- npx -y https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.9.0/ai-architect-mcp-spec.mcpb
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "ai-architect-mcp-spec": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.9.0/ai-architect-mcp-spec.mcpb"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "ai-architect-mcp-spec": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.9.0/ai-architect-mcp-spec.mcpb"
      ]
    }
  }
}
```

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

```json
{
  "servers": {
    "ai-architect-mcp-spec": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.9.0/ai-architect-mcp-spec.mcpb"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "ai-architect-mcp-spec": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.9.0/ai-architect-mcp-spec.mcpb"
      ]
    }
  }
}
```

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

```json
{
  "mcpServers": {
    "ai-architect-mcp-spec": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.9.0/ai-architect-mcp-spec.mcpb"
      ]
    }
  }
}
```

## Tools

- `validate_prd_against_graph`: Checks a spec's code symbols against a supplied code graph.

## FAQ

### Does a model grade the spec?

The structural checks are deterministic with no model in the loop.

### How many tools?

17, per the registry entry.

### How is it packaged?

As an mcpb bundle released on GitHub, over stdio.

### Is it finished?

The site lists AI Architect Spec as still in development.

### Is Ai Architect MCP Spec free?

Yes. Ai Architect MCP Spec is open source, so you can use and self-host it for free.

### How do I add Ai Architect MCP Spec to Claude Code?

Run this command in your terminal: claude mcp add ai-architect-mcp-spec -- npx -y https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.9.0/ai-architect-mcp-spec.mcpb

### What are the best alternatives to Ai Architect MCP Spec?

The closest alternatives to Ai Architect MCP Spec are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.

## Alternatives

- [REA](https://ai-powered.com/mcp/rea): REA is an open-source MCP server for developer tools.
- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [GitHub](https://ai-powered.com/mcp/github-2): GitHub is an open-source MCP server for developer tools.
- [Serena MCP: the IDE for your agent](https://ai-powered.com/mcp/serena-mcp-the-ide-for-your-agent): Serena MCP: the IDE for your agent is an open-source MCP server for developer tools.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [Ouroboros](https://ai-powered.com/mcp/ouroboros): Ouroboros is an open-source MCP server for developer tools.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [Concord MCP](https://ai-powered.com/mcp/concord-mcp): Concord MCP is a freemium MCP server for developer tools.

---
Source: https://ai-powered.com/mcp/ai-architect-mcp-spec
