Skip to main content

Installs the Theseus MCP server into local coding tools

Project description

coding-theorem

Installs the Theseus MCP server into your local coding tools, so your editor can query a knowledge graph of your codebase instead of grepping raw files every time.

Today this means Claude Code. Cursor, Windsurf, and Zed are on the roadmap.

Why

Most coding assistants search text. When you ask about a function, they grep. When you ask how two systems connect, they grep twice and guess. This works until the codebase is big enough that grep returns either nothing useful or too much to read.

Theseus builds and maintains a structured graph of your code, docs, and design decisions. The graph knows which functions call which, where a concept was defined, which commits changed a module and why, and which claims are contested. An MCP server exposes this graph as tools your editor can call directly.

coding-theorem is the thing that wires your editor up to that server. No manual config file editing, no copying UUIDs between windows.

Install

pip install coding-theorem
coding-theorem install

That installs for every detected platform. Target one explicitly:

coding-theorem install --platform claude-code

Use a self-hosted Theseus instance instead of the default:

coding-theorem --mcp-url http://localhost:8001/mcp install --platform claude-code

Restart your editor. The Theseus server should appear in the tool list.

Verify

coding-theorem status

Shows which platforms are configured, which are available to configure, and which MCP URL each one points at.

For Claude Code specifically, a quick manual check:

  1. Restart Claude Code in a repository that Theseus has indexed.
  2. Run a Grep or Glob.
  3. Confirm the hook output includes: coding-theorem: use theseus_code_minimal_context(task) before searching raw files.

If that line appears, the hook is firing and the MCP connection is live.

Uninstall

coding-theorem uninstall

This removes only the managed block and the hook files that coding-theorem created. Your existing config is preserved.

How it works

The installer writes a managed block to each platform's config directory, fenced by markers:

<!-- coding-theorem install: start -->
...generated config...
<!-- coding-theorem install: end -->

Running install twice replaces the block instead of duplicating it. Running uninstall removes only what's inside the markers. Anything you added yourself outside the fences is untouched.

Specifically, coding-theorem install --platform claude-code will:

  • Add a Theseus entry to Claude Code's MCP server list
  • Install a pre-search hook that reminds the model to use graph tools before grep
  • Write assistant hints that describe the available Theseus tools

Run with --dry-run to see exactly what would change without writing anything.

Supported platforms

Platform Status
Claude Code Supported
Cursor Planned
Windsurf Planned
Zed Planned

If you want to add a platform, the pattern is in src/coding_theorem/platforms/claude_code.py. Open an issue first so we can agree on the config shape before you start.

Requirements

  • Python 3.11+
  • A reachable Theseus MCP endpoint (hosted default or self-hosted)
  • One of the supported editors

Hosted vs. self-hosted

The default MCP URL points at the hosted Theseus endpoint. That's fine for trying the tool out, and fine for public repositories.

For private code, self-host. Theseus is open source. Point --mcp-url at your instance and your graph data never leaves your network.

License

MIT. See LICENSE.

Related

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

coding_theorem-0.1.1.tar.gz (10.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

coding_theorem-0.1.1-py3-none-any.whl (13.2 kB view details)

Uploaded Python 3

File details

Details for the file coding_theorem-0.1.1.tar.gz.

File metadata

  • Download URL: coding_theorem-0.1.1.tar.gz
  • Upload date:
  • Size: 10.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for coding_theorem-0.1.1.tar.gz
Algorithm Hash digest
SHA256 fc666146b56dedbcebaf4ce8e455d47960c9c976842c27f3b66969a52fd11cad
MD5 32559dd9ac4d33c13ea95dac96ab5742
BLAKE2b-256 9f56025498d7ec094145feb377d72c3276679fe8df853a52889adc87c56b3b81

See more details on using hashes here.

Provenance

The following attestation bundles were made for coding_theorem-0.1.1.tar.gz:

Publisher: publish.yml on Travis-Gilbert/coding-theorem

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file coding_theorem-0.1.1-py3-none-any.whl.

File metadata

  • Download URL: coding_theorem-0.1.1-py3-none-any.whl
  • Upload date:
  • Size: 13.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for coding_theorem-0.1.1-py3-none-any.whl
Algorithm Hash digest
SHA256 5af78c5909c8c34ce0942b55d5318a4beac0fe4f257da9ae7ee4118a1af7728d
MD5 2fa9f38cc7819ad98745b02e02a7bb35
BLAKE2b-256 5f469a688f61957aaf908365b261c4c54cdc20641b295b9c4d2d341eb6257a26

See more details on using hashes here.

Provenance

The following attestation bundles were made for coding_theorem-0.1.1-py3-none-any.whl:

Publisher: publish.yml on Travis-Gilbert/coding-theorem

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page