codesurface 0.3.0

MCP server that indexes a codebase's public API at startup and serves it via compact tool responses. Pluggable parsers for C#, TypeScript, Python, and more.

codesurface

MCP server that indexes your codebase's public API at startup and serves it via compact tool responses — saving tokens vs reading source files.

Parses source files, extracts public classes/methods/properties/fields/events, and serves them through 5 MCP tools. Works with Claude Code, Cursor, Windsurf, or any MCP-compatible AI tool.

Supported languages: C# (.cs), Python (.py), TypeScript/TSX (.ts, .tsx)

Quick Start

pip install codesurface

Then add to your .mcp.json:

{
  "mcpServers": {
    "codesurface": {
      "command": "codesurface",
      "args": ["--project", "/path/to/your/src"]
    }
  }
}

Point --project at any directory containing supported source files — a Unity Assets/Scripts folder, a .NET src/ tree, a Node.js/React project, a Python package, etc. Languages are auto-detected.

Restart your AI tool and ask: "What methods does MyService have?"

Tools

Tool	Purpose	Example
search	Find APIs by keyword	"MergeService", "BlastBoard", "GridCoord"
get_signature	Exact signature by name or FQN	"TryMerge", "CampGame.Services.IMergeService.TryMerge"
get_class	Full class reference card — all public members	"BlastBoardModel" → all methods/fields/properties
get_stats	Overview of indexed codebase	File count, record counts, namespace breakdown
reindex	Incremental index update (mtime-based)	Only re-parses changed/new/deleted files

Tested On

Project	Language	Files	Records	Time
immich	TypeScript	919	7,957	0.6s
ant-design	TypeScript	2,947	5,452	0.9s
dify	TypeScript	4,903	5,038	1.9s
vscode	TypeScript	6,611	88,293	9.3s
Unity game (private)	C#	129	1,018	0.1s
codesurface (itself)	Python	7	72	<0.1s

Zero duplicate FQNs across all projects.

Benchmarks

Measured against a real Unity game project (129 files, 1,018 API records) across a 10-step cross-cutting research workflow.

Strategy	Total Tokens	vs MCP
MCP (codesurface)	1,021	—
Skilled Agent (Grep + partial Read)	4,453	4.4x more
Naive Agent (Grep + full Read)	11,825	11.6x more

Even with follow-up reads for implementation detail, the hybrid MCP + targeted Read approach uses 54% fewer tokens than a skilled Grep+Read agent.

See workflow-benchmark.md for the full step-by-step analysis.

Setup Details

Claude Code configuration

Add to <project>/.mcp.json:

Using uv (recommended):

{
  "mcpServers": {
    "codesurface": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/codesurface", "codesurface", "--project", "/path/to/your/src"]
    }
  }
}

Using pip install:

{
  "mcpServers": {
    "codesurface": {
      "command": "codesurface",
      "args": ["--project", "/path/to/your/src"]
    }
  }
}

CLAUDE.md snippet (recommended)

Add to your project's CLAUDE.md so the AI knows when to use the tools:

## Codebase API Lookup (codesurface MCP)

Use the `codesurface` MCP tools to look up your project's classes, methods, properties, and fields instead of reading source files.

| When | Tool | Example |
|------|------|---------|
| Searching for an API by keyword | `search` | `search("MergeService")` |
| Need exact method signature | `get_signature` | `get_signature("TryMerge")` |
| Want all members on a class | `get_class` | `get_class("BlastBoardModel")` |
| Overview of indexed codebase | `get_stats` | `get_stats()` |
| After creating/deleting source files | `reindex` | `reindex()` |

Project structure

codesurface/
├── src/codesurface/
│   ├── server.py           # MCP server — 5 tools
│   ├── db.py               # SQLite + FTS5 database layer
│   └── parsers/
│       ├── base.py         # BaseParser ABC
│       ├── csharp.py       # C# parser
│       ├── python_parser.py # Python parser
│       └── typescript.py   # TypeScript/TSX parser
├── pyproject.toml
└── README.md

Troubleshooting

"No codebase indexed"

• Ensure --project points to a directory containing supported source files (.cs, .py, .ts, .tsx)
• The server indexes at startup — check stderr for the "Indexed N records" message

Server won't start

• Check Python version: python --version (needs 3.10+)
• Check mcp[cli] is installed: pip install mcp[cli]

Stale results after editing source files

• Call reindex() — only re-parses files whose modification time changed, fast even on large codebases

---

Contact

[e-mail]

License

MIT