deeprepo-cli 0.2.1

AI project memory for coding tools. Eliminates the cold start tax in every AI session.

deeprepo

Your AI tools forget everything between sessions. deeprepo gives them memory.

Run deeprepo in any project directory and drop into an interactive session — slash commands for project infrastructure, natural language for generating context-rich prompts that copy straight to your clipboard. Paste into Claude Code, Codex, Cursor, or any AI tool. No more re-explaining your codebase every session.

Quickstart

pip install deeprepo-cli
cd your-project/
deeprepo

The first time, deeprepo walks you through API key setup and project initialization. After that, you're in a persistent session:

╭──────────────────────────────────────────────────╮
│  deeprepo v0.2.0                                 │
│  Project: your-project                           │
│  Context: Fresh · 3,008 tokens · 23 files        │
│                                                  │
│  Type /help for commands or ask anything.        │
╰──────────────────────────────────────────────────╯
deeprepo> fix the broken WebSocket connection
OK: Copied prompt (4,200 tokens, 3 files) to clipboard

Type natural language and deeprepo assembles your project context + relevant files + your ask into a clipboard-ready prompt. Paste it into any AI coding tool.

Install

pip install deeprepo-cli

Requires Python 3.11+. One dependency: an OpenRouter API key for the initial project analysis (~$0.50 one-time cost).

What it generates

.deeprepo/
├── config.yaml      # project configuration
├── PROJECT.md       # full project analysis (architecture, patterns, conventions)
├── COLD_START.md    # compressed context prompt optimized for AI consumption
├── SESSION_LOG.md   # running development history
└── SCRATCHPAD.md    # working notes for multi-agent coordination

PROJECT.md is the complete analysis. COLD_START.md is the same content compressed to fit inside a context window — typically under 3,000 tokens. That's what gets included in every prompt you generate.

Interactive TUI

Running deeprepo with no arguments launches the interactive session:

Slash commands map to project infrastructure:

Command	What it does
/init	Analyze project and generate .deeprepo/ context
/context	Copy cold-start prompt to clipboard
/context --format cursor	Write .cursorrules file
/status	Show project context health
/log add <message>	Record development activity
/refresh	Diff-aware context update
/help	List all commands

Natural language builds context-rich prompts:

deeprepo> add rate limiting to the API endpoints
OK: Copied prompt (5,100 tokens, 4 files) to clipboard

deeprepo reads your COLD_START.md, finds files matching your keywords, assembles everything into a structured prompt, and copies it to your clipboard. Paste into any AI tool.

Keybindings: Tab for autocomplete, Ctrl-L to clear, Ctrl-R to refresh context.

CLI mode

Every command also works as a one-shot CLI call:

deeprepo init .                   # Analyze project, generate .deeprepo/
deeprepo context --copy           # Copy cold-start prompt to clipboard
deeprepo context --format cursor  # Write .cursorrules file
deeprepo refresh --full           # Full re-analysis
deeprepo status                   # Context health dashboard
deeprepo log "fixed auth bug"     # Record session activity
deeprepo tui /path/to/project     # Launch TUI for a specific path

How it works

deeprepo uses a Recursive Language Model (RLM) pattern — a root LLM operates in a Python REPL loop, exploring your codebase and dispatching focused analysis tasks to cheap sub-LLM workers.

┌─────────────────────────────────────────────────┐
│  Root Orchestrator (Claude Sonnet 4.5)          │
│                                                  │
│  Sees: file tree, metadata, sizes               │
│  Does NOT see: actual file contents              │
│                                                  │
│  Writes Python → explores codebase → dispatches  │
│  analysis tasks → synthesizes findings           │
└──────────────────┬──────────────────────────────┘
                   │ llm_query() / llm_batch()
                   ▼
┌─────────────────────────────────────────────────┐
│  Sub-LLM Workers (MiniMax M2.5)                 │
│                                                  │
│  Each worker gets ONE focused task:              │
│  "Analyze auth.py for security issues"           │
│  "Map the data flow in this module"              │
│                                                  │
│  Cost: ~$0.002 per file analysis                 │
└─────────────────────────────────────────────────┘

The root model never sees raw file contents — it writes Python code to explore, dispatches individual files to cheap workers, and synthesizes findings across turns. All sub-LLM calls go through OpenRouter, so one API key gives you access to every model.

Benchmarks

The RLM engine was benchmarked on FastAPI (47 files) — same codebase, same task, two approaches:

Metric	Sonnet RLM	Opus Baseline
Total cost	$0.46	$0.99
Sub-LLM calls	13	N/A
Sub-LLM cost	$0.02	N/A
Files covered	47/47 (100%)	42/47 (89%)

The 5 files the baseline missed account for 74% of FastAPI's codebase by size — including the core FastAPI class, routing engine, and dependency injection system. The coverage gap widens with scale: the baseline dropped to 48% on a 289-file codebase while the RLM maintained 100%.

Full data in BENCHMARK_RESULTS.md.

Configuration

# Required: OpenRouter key for sub-LLM workers + alternative root models
export OPENROUTER_API_KEY=sk-or-...

# Optional: Anthropic key for Claude root models (falls back to OpenRouter)
export ANTHROPIC_API_KEY=sk-ant-...

Or let deeprepo set it up for you — the interactive onboarding saves your key to ~/.deeprepo/config.yaml on first run.

After deeprepo init, project settings live in .deeprepo/config.yaml:

root_model: anthropic/claude-sonnet-4-5
sub_model: minimax/minimax-m2.5
context_max_tokens: 3000
include_tech_debt: true

Short aliases work on the CLI: --root-model sonnet, --root-model opus, --sub-model minimax.

Roadmap

• Teams — named multi-agent compositions for different analysis workflows
• More --format targets — Windsurf, Aider, generic system prompt
• Embeddings-based file matching — smarter relevant file detection in prompt builder

License

MIT