| Crates.io | acp-mcp |
| lib.rs | acp-mcp |
| version | 0.2.0 |
| created_at | 2025-12-29 16:43:35.036068+00 |
| updated_at | 2026-01-02 15:29:15.168529+00 |
| description | ACP MCP Server - Model Context Protocol for AI tools |
| homepage | |
| repository | https://github.com/acp-protocol/acp-mcp |
| max_upload_size | |
| id | 2010904 |
| size | 263,914 |
Model Context Protocol (MCP) server for the AI Context Protocol.
# Via cargo
cargo install acp-mcp
# Or via ACP CLI
acp install mcp
The MCP server runs over stdio for integration with AI tools like Claude Desktop.
# Run MCP server
acp-mcp
# With custom project root
acp-mcp -C /path/to/project
# With debug logging
acp-mcp --log-level debug
Add to your Claude Desktop configuration (claude_desktop_config.json):
{
"mcpServers": {
"acp": {
"command": "acp-mcp",
"args": ["-C", "/path/to/your/project"]
}
}
}
| Tool | Description |
|---|---|
acp_get_architecture |
Get project overview and structure |
acp_get_file_context |
Get file details with relationships |
acp_get_symbol_context |
Get symbol analysis with call graphs |
acp_get_domain_files |
Query files by domain |
acp_check_constraints |
Verify constraint compliance |
acp_get_hotpaths |
Find critical/frequently-called symbols |
acp_expand_variable |
Resolve variable values |
acp_generate_primer |
Generate optimized AI context |
The MCP server reads ACP files from the project root:
.acp/acp.cache.json - Indexed cache (required).acp/acp.vars.json - Variables (optional).acp.config.json - Configuration (optional)Generate these with the ACP CLI:
acp index
MIT