codegraph-cli 2.1.3

AI-powered code intelligence CLI with multi-agent analysis, impact graphs, and conversational coding.

CodeGraph CLI

Code intelligence from the terminal. Semantic search, impact analysis, multi-agent code generation, and conversational coding — all backed by your choice of LLM.

---

Overview

CodeGraph CLI (cg) parses your codebase into a semantic graph, then exposes that graph through search, impact analysis, visualization, and a conversational interface. It supports six LLM providers and optionally runs a CrewAI multi-agent system that can read, write, patch, and rollback files autonomously.

Core capabilities:

• Semantic Search — find code by meaning, not string matching
• Impact Analysis — trace multi-hop dependencies before making changes
• Graph Visualization — interactive HTML and Graphviz DOT exports
• Browser-Based Explorer — visual code navigation with Mermaid diagrams and AI explanations
• Conversational Chat — natural language coding sessions with RAG context
• Multi-Agent System — CrewAI-powered agents for code generation, refactoring, and analysis
• DOCX Export — generate professional project documentation with architecture diagrams
• Auto Onboarding — AI-generated README from your code graph
• File Rollback — automatic backups before every file modification

---

Installation

pip install codegraph-cli

With neural embedding models (semantic code search):

pip install codegraph-cli[embeddings]

With CrewAI multi-agent support:

pip install codegraph-cli[crew]

Everything:

pip install codegraph-cli[all]

For development:

git clone https://github.com/al1-nasir/codegraph-cli.git
cd codegraph-cli
pip install -e ".[dev]"

---

Quick Start

1. Configure your LLM provider

cg config setup

This runs an interactive wizard that writes configuration to ~/.codegraph/config.toml. Alternatively, switch providers directly:

cg config set-llm openrouter
cg config set-llm groq
cg config set-llm ollama

2. Index a project

cg project index /path/to/project --name myproject

This parses the source tree using tree-sitter, builds a dependency graph in SQLite, and generates embeddings for semantic search.

3. Use it

cg analyze search "authentication logic"
cg analyze impact UserService --hops 3
cg analyze graph process_payment --depth 2
cg chat start
cg chat start --crew       # multi-agent mode
cg explore open            # browser-based code explorer
cg onboard                 # auto-generate project README
cg export docx             # export documentation to DOCX

---

Supported LLM Providers

Provider	Type	Configuration
Ollama	Local, free	cg config set-llm ollama
Groq	Cloud, free tier	cg config set-llm groq
OpenAI	Cloud	cg config set-llm openai
Anthropic	Cloud	cg config set-llm anthropic
Gemini	Cloud	cg config set-llm gemini
OpenRouter	Cloud, multi-model	cg config set-llm openrouter

All configuration is stored in ~/.codegraph/config.toml. No environment variables required.

cg config show-llm        # view current provider, model, and endpoint
cg config unset-llm       # reset to defaults

---

Embedding Models

CodeGraph supports configurable embedding models for semantic code search. Choose based on your hardware and quality needs:

Model	Download	Dim	Quality	Command
hash	0 bytes	256	Keyword-only	cg config set-embedding hash
minilm	~80 MB	384	Decent	cg config set-embedding minilm
bge-base	~440 MB	768	Good	cg config set-embedding bge-base
jina-code	~550 MB	768	Code-aware	cg config set-embedding jina-code
qodo-1.5b	~6.2 GB	1536	Best	cg config set-embedding qodo-1.5b

The default is hash (zero-dependency, no download). Neural models require the [embeddings] extra and are downloaded on first use from HuggingFace.

cg config set-embedding jina-code    # switch to a neural model
cg config show-embedding             # view current model and all options
cg config unset-embedding            # reset to hash default

After changing the embedding model, re-index your project:

cg index /path/to/project

---

Commands

CodeGraph CLI organizes commands into logical groups:

cg config    — LLM, embedding, and setup configuration
cg project   — Index, load, and manage project memories
cg analyze   — Search, impact, graph, and RAG context
cg chat      — Interactive chat with AI agents
cg explore   — Visual code explorer in browser
cg export    — Export project documentation
cg onboard   — Auto-generate project README

Configuration (cg config)

cg config setup                      # interactive LLM setup wizard
cg config set-llm openrouter         # switch LLM provider
cg config unset-llm                  # reset LLM config to defaults
cg config show-llm                   # show current LLM settings
cg config set-embedding jina-code    # switch embedding model
cg config unset-embedding            # reset to hash default
cg config show-embedding             # show current embedding model

Project Management (cg project)

cg project index <path> [--name NAME]  # parse and index a codebase
cg project list                        # list all indexed projects
cg project load <name>                 # switch active project
cg project current                     # print active project name
cg project delete <name>               # remove a project index
cg project merge <src> <dst>           # merge two project graphs
cg project unload                      # unload without deleting
cg project init                        # quickstart wizard
cg project watch                       # auto-reindex on file changes

Code Analysis (cg analyze)

cg analyze search <query> [--top-k N]  # semantic search across the graph
cg analyze impact <symbol> [--hops N]  # multi-hop dependency impact analysis
cg analyze graph <symbol> [--depth N]  # ASCII dependency graph
cg analyze export-graph --format html  # interactive vis.js visualization
cg analyze export-graph --format dot   # Graphviz DOT format
cg analyze rag-context <query>         # raw RAG retrieval for debugging
cg analyze tree [--full]               # directory tree of indexed project
cg analyze sync [--dry-run]            # incremental index sync
cg analyze health                      # project health dashboard

Interactive Chat (cg chat)

cg chat start                          # start or resume a session
cg chat start --new                    # force a new session
cg chat start --crew                   # multi-agent mode (CrewAI)
cg chat start -s <id>                  # resume a specific session
cg chat list                           # list all sessions
cg chat delete <id>                    # delete a session
cg chat export <id> --format markdown  # export session to file

In-chat commands:

Command	Mode	Description
/help	Both	Show available commands
/clear	Both	Clear conversation history
/new	Both	Start a fresh session
/exit	Both	Save and exit
/apply	Standard	Apply pending code proposal
/preview	Standard	Preview pending file changes
/backups	Crew	List all file backups
/rollback <file>	Crew	Restore a file from backup
/undo <file>	Crew	Alias for rollback

Visual Explorer (cg explore)

cg explore open                        # launch browser-based code explorer
cg explore open --port 9000            # use custom port

Opens a local web UI with directory tree navigation, syntax-highlighted code, AI explanations, dependency graphs, and Mermaid diagrams.

Documentation Export (cg export)

cg export docx                         # basic DOCX with structure + diagrams
cg export docx --enhanced              # add AI-powered explanations
cg export docx --include-code          # include source code listings
cg export docx --enhanced --depth files --include-code  # full export

Auto Onboarding

cg onboard                             # print AI-generated README to stdout
cg onboard --save                      # save as ONBOARD.md in project dir
cg onboard -o README.md                # save to specific file
cg onboard --no-llm                    # template-only, no LLM call

---

Multi-Agent System

When you run cg chat start --crew, CodeGraph launches a CrewAI pipeline with four specialized agents:

Agent	Role	Tools
Project Coordinator	Routes tasks to the right specialist	Delegation only
File System Engineer	File I/O, directory traversal, backups	list_directory, read_file, write_file, patch_file, delete_file, rollback_file, file_tree
Senior Software Developer	Code generation, refactoring, bug fixes	All tools (file ops + code analysis)
Code Intelligence Analyst	Search, dependency tracing, explanations	search_code, grep, project_summary, read_file

Every file modification automatically creates a timestamped backup in ~/.codegraph/backups/. Files can be rolled back to any previous state via /rollback or cg v2 rollback.

---

Architecture

CLI Layer (Typer + Rich)
    |
    +-- config ─────────> ConfigManager (TOML)
    |
    +-- project ────────> MCPOrchestrator ──> GraphStore (SQLite)
    |                         |                    |
    |                         +-- Parser           +-- VectorStore (LanceDB)
    |                         |   (tree-sitter)    |
    |                         +-- RAGRetriever     +-- Embeddings (configurable)
    |                         +-- LLM Adapter          hash | minilm | bge-base
    |                                                  jina-code | qodo-1.5b
    +-- analyze ────────> Search, Impact, Graph, Tree, Sync, Health
    |
    +-- chat ───────────> ChatAgent (standard mode)
    |                     CrewChatAgent (--crew mode)
    |                         +-- Coordinator Agent
    |                         +-- File System Agent ──> 8 file operation tools
    |                         +-- Code Gen Agent ─────> all 11 tools
    |                         +-- Code Analysis Agent > 3 search/analysis tools
    |
    +-- explore ────────> Starlette + Uvicorn (browser UI)
    |
    +-- export ─────────> DOCX generator + Mermaid diagrams
    |
    +-- onboard ────────> AI README generation from code graph

Embeddings: Five models available via cg config set-embedding. Hash (default, zero-dependency) through Qodo-Embed-1-1.5B (best quality, 6 GB). Neural models use raw transformers + torch — no sentence-transformers overhead. Models are cached in ~/.codegraph/models/.

Parser: tree-sitter grammars for Python, JavaScript, and TypeScript. Extracts modules, classes, functions, imports, and call relationships into a directed graph.

Storage: SQLite for the code graph (nodes + edges), LanceDB for vector embeddings. All data stored under ~/.codegraph/.

LLM Adapter: Unified interface across six providers. For CrewAI, models are routed through LiteLLM. Configuration is read exclusively from ~/.codegraph/config.toml.

---

Project Structure

codegraph_cli/
    cli.py               # main Typer application, command wiring
    cli_groups.py         # command group definitions (config, project, analyze)
    cli_chat.py           # interactive chat REPL with Rich output
    cli_setup.py          # setup wizard, set-llm, set-embedding
    cli_explore.py        # browser-based visual code explorer (Starlette)
    cli_export.py         # DOCX export with Mermaid diagrams
    cli_onboard.py        # AI-generated project README
    cli_health.py         # project health dashboard
    cli_quickstart.py     # quickstart / init wizard
    cli_watch.py          # auto-reindex on file changes
    config.py             # loads config from TOML
    config_manager.py     # TOML read/write, provider and embedding config
    llm.py                # multi-provider LLM adapter
    parser.py             # tree-sitter AST parsing (Python, JS, TS)
    storage.py            # SQLite graph store
    embeddings.py         # configurable embedding engine (5 models)
    rag.py                # RAG retriever
    vector_store.py       # LanceDB vector store
    orchestrator.py       # coordinates parsing, search, impact
    graph_export.py       # DOT and HTML graph export
    project_context.py    # unified file access layer
    crew_tools.py         # 11 CrewAI tools (file ops + analysis)
    crew_agents.py        # 4 specialized CrewAI agents
    crew_chat.py          # CrewAI orchestrator with rollback
    chat_agent.py         # standard chat agent
    chat_session.py       # session persistence
    models.py             # core data models
    templates/
        graph_interactive.html  # vis.js graph template

---

Development

git clone https://github.com/al1-nasir/codegraph-cli.git
cd codegraph-cli
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev,crew,embeddings]"
pytest

---

License

MIT. See LICENSE.