# archeus

> The memory and workspace layer for AI coding agents. Persistent per-project memory, every session you have ever had, and control over what the next one costs. Works with Claude Code today: a searchable session archive, an interactive architecture graph, MCP management and per-project launch control. Pure Python standard library, zero runtime dependencies, MIT.

archeus is a Python command-line and desktop tool that wraps the Claude Code CLI. It is MIT-licensed, runs on the Python standard library alone, and needs no API key — it uses the Claude Code authentication you already have. Install with `pipx install archeus`. Source: https://github.com/babarmuhammad/archeus. Package: https://pypi.org/project/archeus/. Documentation: https://docs.claudectl.space. Author: Babar Muhammad Anas.

Where this is going, as a long-term goal and not as a shipped feature: provider-neutral memory, then an open harness of its own. Claude Code is the first surface, not the boundary. Today Claude Code is the only agent archeus drives, and nothing else is supported.

Disambiguation: two independent open-source projects use the name "archeus". This one is the Python memory and workspace layer for AI coding agents described above, published on PyPI at https://pypi.org/project/archeus/. The other is a Rust agent orchestrator by a different author, published on crates.io at https://crates.io/crates/archeus. They are unrelated, and neither is affiliated with Anthropic.

## Documentation

- [Documentation](https://docs.claudectl.space): The full reference manual: installation, every screen, the memory model, hooks, MCP and the HTTP API.
- [Getting started](https://docs.claudectl.space/getting-started/): Install, open a project, and launch the first session.

## Pages

- [The memory and workspace layer for AI coding agents.](https://claudectl.space/): The memory and workspace layer for AI coding agents.
- [Features](https://claudectl.space/features): Everything archeus does — sessions and search, project memory, the architecture graph, Plan to Execute, multiple accounts, MCP servers, agents, hooks, the status line and the desktop GUI.
- [Download archeus](https://claudectl.space/download): Every way to install archeus — pipx, pip, a GitHub release, a git checkout or the Claude Code plugin.
- [Architecture](https://claudectl.space/architecture): How archeus is built — the connections engine, the memory graph, the job runner, the local HTTP API behind the GUI, and the interactive dependency graph it draws of your own project.
- [Community](https://claudectl.space/community): Where to report an archeus bug, ask a question, propose a feature and read the code of conduct.
- [About archeus](https://claudectl.space/about): Who builds archeus, the MIT licence it ships under, and the Claude Code ecosystem projects its memory, review, skills and model routing are built on.
- [Contributing to archeus](https://claudectl.space/contributing): How to work on archeus — the development setup, the test suite as a contract, the lint gate, and what makes a change land.
- [Blog](https://claudectl.space/blog): Long-form articles on Claude Code memory, context cost, sessions and multi-account setups.
- [FAQ](https://claudectl.space/faq): Direct answers to the questions people actually ask about archeus and Claude Code.
- [Changelog](https://claudectl.space/changelog): Every release, newest first, generated from the repository CHANGELOG.
- [Code of conduct](https://claudectl.space/code-of-conduct): The behaviour expected of contributors.

## Blog

- [How to cut Claude Code token costs: the 250-token index pattern](https://claudectl.space/blog/cut-claude-code-token-costs): Context files ride along on every message, so their size is a permanent per-turn tax.
- [Managing Claude Code sessions on Windows](https://claudectl.space/blog/managing-claude-code-sessions-on-windows): Claude Code writes every session to a JSONL file in an encoded folder and gives you no way to browse them.
- [Running multiple Claude accounts without losing your place](https://claudectl.space/blog/multiple-claude-accounts): Claude Code picks its account from CLAUDE_CONFIG_DIR.
- [Plan on an expensive model, execute on a free one](https://claudectl.space/blog/plan-expensive-execute-cheap): Run the expensive reasoning once, then hand the approved plan to a cheap or free model to build.
- [Reading Claude Code's own state: sessions, plugins, checkpoints, settings](https://claudectl.space/blog/reading-claude-codes-own-state): Claude Code keeps transcripts, checkpoints, plugin caches and settings on disk in formats it never documented.
- [What belongs in CLAUDE.md, and what belongs in path-scoped rules](https://claudectl.space/blog/what-belongs-in-claude-md): CLAUDE.md is read on every turn; path-scoped rules cost nothing until Claude opens a matching file.
- [Why Claude Code forgets your project between sessions, and what to do about it](https://claudectl.space/blog/why-claude-code-forgets-your-project): Claude Code starts every session with an empty context window.

## Optional

- [llms-full.txt](https://claudectl.space/llms-full.txt): Every page above, the FAQ and the changelog as one plain-text file.
- [Source repository](https://github.com/babarmuhammad/archeus): The code, issues and releases.
- [PyPI package](https://pypi.org/project/archeus/): Published wheels; install with pipx install archeus.