codesavant 1.2.9 → 2.3.0

package/README.md

@@ -1,845 +1,104 @@
 # CodeSavant
 
-<div align="center">
+AI-powered coding assistant CLI with multi-provider support. A feature-complete alternative to Claude Code CLI that works with multiple LLM providers.
 
-```
- ██████╗ ██████╗ ██████╗ ███████╗███████╗ █████╗ ██╗   ██╗ █████╗ ███╗   ██╗████████╗
-██╔════╝██╔═══██╗██╔══██╗██╔════╝██╔════╝██╔══██╗██║   ██║██╔══██╗████╗  ██║╚══██╔══╝
-██║     ██║   ██║██║  ██║█████╗  ███████╗███████║██║   ██║███████║██╔██╗ ██║   ██║
-██║     ██║   ██║██║  ██║██╔══╝  ╚════██║██╔══██║╚██╗ ██╔╝██╔══██║██║╚██╗██║   ██║
-╚██████╗╚██████╔╝██████╔╝███████╗███████║██║  ██║ ╚████╔╝ ██║  ██║██║ ╚████║   ██║
- ╚═════╝ ╚═════╝ ╚═════╝ ╚══════╝╚══════╝╚═╝  ╚═╝  ╚═══╝  ╚═╝  ╚═╝╚═╝  ╚═══╝   ╚═╝
-```
-
-**Multi-Provider AI Coding Assistant for Your Terminal**
-
-[![npm version](https://badge.fury.io/js/codesavant.svg)](https://www.npmjs.com/package/codesavant)
-[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![Node.js Version](https://img.shields.io/node/v/codesavant.svg)](https://nodejs.org)
-
-</div>
-
----
-
-## What is CodeSavant?
-
-CodeSavant is a powerful AI coding assistant that runs directly in your terminal. Unlike cloud-based solutions, CodeSavant gives you the flexibility to choose from multiple AI providers (DeepSeek, Groq, Gemini, OpenAI, Anthropic, or local Ollama) while keeping your code and conversations private.
-
-**Key Benefits:**
-- **Cost Effective**: Use providers like DeepSeek at $0.14/1M tokens vs $15/1M for GPT-4
-- **Provider Choice**: Switch between 7+ AI providers based on your needs
-- **Privacy First**: Run locally with Ollama, or use any cloud provider
-- **Full Tooling**: Read, write, edit files, run commands, search codebases
-- **Session Persistence**: Continue conversations across terminal sessions
-- **Extensible**: Custom skills, MCP servers, and subagents
+## Features
 
----
-
-## Table of Contents
-
-- [Installation](#installation)
-- [Quick Start](#quick-start)
-- [Supported Providers](#supported-providers)
-- [Usage](#usage)
-  - [Interactive Mode](#interactive-mode)
-  - [REPL Commands](#repl-commands)
-  - [Multi-line Input](#multi-line-input)
-  - [Keyboard Shortcuts](#keyboard-shortcuts)
-- [Tools](#tools)
-- [Subagents](#subagents)
-- [Skills System](#skills-system)
-- [Permission System](#permission-system)
-- [Session Management](#session-management)
-- [Configuration](#configuration)
-- [Project Memory (SAVANT.md)](#project-memory-savantmd)
-- [MCP Integration](#mcp-integration)
-- [Live Preview](#live-preview)
-- [CI/CD Fix Mode](#cicd-fix-mode)
-- [Checkpoints & Recovery](#checkpoints--recovery)
-- [Enterprise Features](#enterprise-features)
-- [Debug & Troubleshooting](#debug--troubleshooting)
-- [CLI Reference](#cli-reference)
-- [Environment Variables](#environment-variables)
-- [License](#license)
-
----
+- **Multi-Provider Support**: Anthropic, OpenAI, Gemini, Bedrock, Groq, DeepSeek, Ollama
+- **Full Tool Suite**: Read, Write, Edit, Bash, Glob, Grep, LSP, MultiEdit, LS, NotebookRead
+- **MCP Integration**: Connect to Model Context Protocol servers
+- **Hooks System**: 40+ hook events for customization
+- **Session Management**: Resume previous conversations
+- **Plan Mode**: Read-only exploration mode
+- **Subagent System**: Spawn specialized agents for complex tasks
 
 ## Installation
 
-### Prerequisites
-
-- Node.js 18.0.0 or higher
-- At least one AI provider API key (or Ollama installed locally)
-- [ripgrep](https://github.com/BurntSushi/ripgrep) (recommended for fast code search)
-
-### Install via npm
-
 ```bash
-npm install -g codesavant
-```
+# Using bun (recommended)
+bun add -g codesavant
 
-### Verify Installation
-
-```bash
-codesavant --version
-codesavant doctor  # Run diagnostic checks
-```
-
----
-
-## Quick Start
-
-### 1. Set up an API key
-
-Choose at least one provider and set the API key:
-
-```bash
-# DeepSeek (Recommended - Best cost/performance ratio)
-export DEEPSEEK_API_KEY="your-api-key"
-
-# Groq (Fastest response times)
-export GROQ_API_KEY="your-api-key"
-
-# Google Gemini (Large context window - 1M tokens)
-export GEMINI_API_KEY="your-api-key"
-
-# OpenAI
-export OPENAI_API_KEY="your-api-key"
-
-# Anthropic
-export ANTHROPIC_API_KEY="your-api-key"
-
-# Or use Ollama locally (no API key needed)
-# Just ensure Ollama is running: ollama serve
-```
-
-### 2. Start CodeSavant
-
-```bash
-codesavant
-```
-
-### 3. Start Coding!
-
-```
-> Explain the structure of this codebase
-
-> Create a function to validate email addresses
-
-> Fix the bug in src/utils.ts
-
-> Write tests for the authentication module
+# Using npm
+npm install -g codesavant
 ```
 
----
-
-## Supported Providers
-
-| Provider | Cost (per 1M tokens) | Speed | Best For | API Key Variable |
-|----------|---------------------|-------|----------|------------------|
-| **DeepSeek** | $0.14 in / $0.28 out | Medium | Primary coding tasks | `DEEPSEEK_API_KEY` |
-| **Groq** | $0.59 in / $0.79 out | Fastest | Quick queries, chat | `GROQ_API_KEY` |
-| **Gemini** | $0.15 in / $0.60 out | Fast | Large context tasks | `GEMINI_API_KEY` |
-| **OpenAI** | $5.00 in / $15.00 out | Medium | Complex reasoning | `OPENAI_API_KEY` |
-| **Anthropic** | $3.00 in / $15.00 out | Medium | Code review, analysis | `ANTHROPIC_API_KEY` |
-| **Vertex AI** | Varies | Medium | Enterprise/GCP users | `GOOGLE_APPLICATION_CREDENTIALS` |
-| **Ollama** | Free (local) | Variable | Offline, privacy | None (local) |
-
-**Estimated Monthly Cost**: $2-10 for typical individual usage with DeepSeek/Groq.
-
-### Smart Routing
+## Requirements
 
-CodeSavant automatically routes requests to the best available provider based on:
-- Task complexity
-- Provider availability
-- Cost optimization
-- Response speed requirements
-
----
+- [Bun](https://bun.sh) >= 1.0.0
 
 ## Usage
 
-### Interactive Mode
-
 ```bash
-# Start a new session
+# Interactive mode
 codesavant
 
-# Continue the last session
-codesavant -c
-codesavant --continue
-
-# Resume a specific session
-codesavant -r <session-id>
-codesavant --resume <session-id>
-
-# Start with an initial prompt
+# Single prompt
 codesavant "Explain this codebase"
 
-# Use a specific provider
-codesavant --provider deepseek
-codesavant --provider groq
-
-# Use a specific model
-codesavant --model gpt-4o
-
-# Use classic terminal UI (no Ink)
-codesavant --classic
-```
-
-### REPL Commands
-
-Inside the CodeSavant REPL, use these slash commands:
-
-| Command | Description |
-|---------|-------------|
-| `/help` | Show all available commands |
-| `/clear` | Clear conversation history |
-| `/history` | Show conversation history |
-| `/session` | Show current session ID |
-| `/sessions` | List all saved sessions |
-| `/providers` | Show available AI providers |
-| `/stats` | Show token usage and cost statistics |
-| `/compact` | Manually compact conversation context |
-| `/export [format]` | Export conversation (markdown, json, html) |
-| `/config` | Show current configuration |
-| `/doctor` | Diagnose configuration issues |
-| `/init` | Initialize CodeSavant in current project |
-| `/preview` | Manage live preview servers |
-| `/status` | Show session status and statistics |
-| `/debug` | Debug logging controls |
-| `/exit` or `Ctrl+D` | Exit CodeSavant |
-
-### Multi-line Input
-
-Use triple quotes `"""` to enter multi-line input:
-
-```
-> """
-Create a React component that:
-- Displays a list of users
-- Supports pagination
-- Has a search filter
-- Shows loading states
-"""
-```
-
-Press Enter after the closing `"""` to submit.
-
-### Keyboard Shortcuts
-
-| Shortcut | Action |
-|----------|--------|
-| `Tab` | Autocomplete commands |
-| `↑/↓` | Navigate command history |
-| `Ctrl+R` | Search command history |
-| `Ctrl+C` | Cancel current operation |
-| `Ctrl+D` | Exit CodeSavant |
-| `Shift+Tab` | Cycle permission modes |
-| `Esc Esc` | Rewind to last checkpoint |
-| `"""` | Enter multi-line mode |
-
----
-
-## Tools
-
-CodeSavant has 17 built-in tools that the AI can use:
-
-### File Operations
-
-| Tool | Description |
-|------|-------------|
-| **read_file** | Read file contents with line numbers |
-| **write_file** | Create or overwrite files |
-| **edit_file** | Make precise edits with search/replace |
-| **multi_edit** | Apply multiple atomic edits across files |
-| **list_directory** | List directory contents |
-
-### Search Operations
-
-| Tool | Description |
-|------|-------------|
-| **glob** | Find files using glob patterns |
-| **grep** | Search file contents with regex (ripgrep) |
-
-### System Operations
-
-| Tool | Description |
-|------|-------------|
-| **bash** | Execute shell commands (120s timeout) |
-| **notebook_edit** | Edit Jupyter notebooks (.ipynb) |
-| **screenshot** | Capture screen screenshots |
-| **computer_use** | Control mouse and keyboard |
-| **preview** | Manage live preview of web projects |
-
-### Network Operations
-
-| Tool | Description |
-|------|-------------|
-| **web_fetch** | Fetch URLs and convert HTML to Markdown |
-| **web_search** | Search the web |
-
-### AI Operations
-
-| Tool | Description |
-|------|-------------|
-| **spawn_subagent** | Delegate tasks to specialized agents |
-| **todo_write** | Manage structured task lists |
-| **ask_user_question** | Ask user questions with choices |
-
-### Examples
-
-```
-> Read the package.json file
-
-> Find all TypeScript files in src/
-
-> Search for "TODO" comments in the codebase
-
-> Run npm test
-
-> Create a new component at src/components/Button.tsx
-```
-
----
-
-## Subagents
-
-Specialized AI agents that can be spawned for focused tasks:
-
-| Subagent | Description | Best For |
-|----------|-------------|----------|
-| **explore** | Codebase exploration specialist | Understanding code structure, finding patterns |
-| **plan** | Research and planning agent | Creating implementation plans, architecture |
-| **bash** | Command execution specialist | Git operations, system tasks |
-| **general-purpose** | Multi-tool agent | Complex multi-step tasks |
-| **test-runner** | Test execution specialist | Running and analyzing tests |
-| **build-validator** | Build verification | Fixing compilation errors |
-| **code-guide** | Documentation lookup | CodeSavant feature help |
-
-### Using Subagents
-
-The AI automatically uses subagents for appropriate tasks, or you can request them:
-
-```
-> Explore this codebase and explain the architecture
-
-> Plan how to implement user authentication
-
-> Run the tests and fix any failures
-```
-
----
-
-## Skills System
-
-Skills are pre-built or custom AI behaviors for common tasks.
-
-### Built-in Skills
-
-| Skill | Description |
-|-------|-------------|
-| `/commit` | Generate a git commit message |
-| `/review` | Review code for issues |
-| `/test` | Generate tests for code |
-| `/explain` | Explain how code works |
-| `/fix` | Debug and fix issues |
-
-### Using Skills
-
-```
-> /commit
-
-> /review src/auth.ts
-
-> /test src/utils/validation.ts
-```
-
-### Creating Custom Skills
-
-Create a `SKILL.md` file in `~/.codesavant/skills/` or `.codesavant/skills/`:
-
-```yaml
----
-name: document
-description: Generate documentation for code
-triggers: ['document', 'docs', 'jsdoc']
-allowedTools: ['read_file', 'write_file', 'edit_file']
-permissionMode: acceptEdits
-maxTurns: 10
-author: Your Name
-version: 1.0
----
-
-# Documentation Generator
-
-Generate comprehensive documentation for the specified code...
-```
-
----
-
-## Permission System
-
-CodeSavant protects your system with a granular permission system.
-
-### Permission Modes
-
-| Mode | Description |
-|------|-------------|
-| `default` | Ask for dangerous operations, auto-allow safe ones |
-| `plan` | Read-only mode, no file modifications |
-| `acceptEdits` | Auto-approve file edits |
-| `strict` | Ask for all operations |
-| `yolo` | Auto-approve everything (dangerous!) |
-
-### Cycling Modes
-
-Press `Shift+Tab` in the REPL to cycle through permission modes, or use:
-
-```bash
-codesavant --mode plan
-codesavant --mode strict
-```
-
-### Interactive Prompts
-
-When prompted for permission:
-- `y` or `1` - Allow this once
-- `n` or `2` - Deny this once
-- `a` or `3` - Always allow (this session)
-- `v` or `4` - Never allow (this session)
+# Print mode (non-interactive)
+codesavant -p "List all TypeScript files"
 
-### Auto-Denied Operations
+# Pipe mode
+echo "Summarize this file" | codesavant --pipe
 
-- `rm -rf /` and similar destructive commands
-- `sudo` commands
-- Writing to `.env` files
-- Fork bombs and shell injection
-
----
-
-## Session Management
-
-Sessions persist conversations across terminal sessions.
-
-### Session Commands
-
-```bash
-# List all sessions
-codesavant sessions -l
-
-# Delete a session
-codesavant sessions -d <session-id>
-
-# Clean up old sessions (default: 30 days)
-codesavant sessions --cleanup [days]
-
-# Continue most recent session
-codesavant -c
-
-# Resume specific session
+# Resume session
 codesavant -r <session-id>
-```
-
-### Session Features
-
-- **Atomic Writes**: Prevents data corruption
-- **Cost Tracking**: Per-message token usage and cost
-- **Checkpoints**: Save/restore conversation state
-- **Export**: Export to Markdown, JSON, or HTML
-
----
-
-## Configuration
-
-### Configuration File
-
-Location: `~/.codesavant/config.json`
-
-```json
-{
-  "defaultProvider": "deepseek",
-  "maxTokens": 8192,
-  "routingStrategy": "auto",
-  "providers": [
-    {
-      "name": "deepseek",
-      "model": "deepseek-chat",
-      "apiKey": "sk-...",
-      "costPerMToken": {"input": 0.14, "output": 0.28}
-    }
-  ],
-  "permissions": []
-}
-```
-
-### Configuration Commands
-
-```bash
-# View all configuration
-codesavant config --show
-
-# View available providers
-codesavant config --providers
-
-# Set a value
-codesavant config set defaultProvider deepseek
-```
-
-### Directory Structure
-
-```
-~/.codesavant/
-├── config.json      # Global configuration
-├── mcp.json         # MCP server configurations
-├── SAVANT.md        # Global memory
-├── sessions/        # Saved sessions
-├── checkpoints/     # Conversation checkpoints
-├── audit/           # Permission audit logs
-└── skills/          # Custom skills
-```
-
----
-
-## Project Memory (SAVANT.md)
-
-CodeSavant uses a three-level memory hierarchy:
-
-| Level | File | Purpose |
-|-------|------|---------|
-| **User** | `~/.codesavant/SAVANT.md` | Global preferences |
-| **Project** | `./SAVANT.md` | Project-specific context |
-| **Local** | `./SAVANT.local.md` | Personal overrides (gitignored) |
-
-### Example SAVANT.md
-
-```markdown
-# Project: MyApp
-
-## Overview
-E-commerce platform built with Next.js and PostgreSQL.
-
-## Tech Stack
-- Frontend: Next.js 14, React 18, TailwindCSS
-- Backend: Node.js, Express, Prisma ORM
-- Database: PostgreSQL
-- Testing: Jest, Playwright
-
-## Project Structure
-- `/src/app` - Next.js app router pages
-- `/src/components` - Reusable React components
-- `/src/lib` - Utility functions
-- `/prisma` - Database schema
-
-## Coding Conventions
-- Use TypeScript strict mode
-- Prefer functional components
-- Run `npm run lint` before committing
-
-## Common Commands
-- `npm run dev` - Start dev server
-- `npm run build` - Build for production
-- `npm run test` - Run tests
-```
-
----
-
-## MCP Integration
-
-CodeSavant supports the Model Context Protocol (MCP) for extending functionality.
-
-### Configure MCP Servers
-
-Create `~/.codesavant/mcp.json`:
-
-```json
-{
-  "servers": {
-    "filesystem": {
-      "transport": "stdio",
-      "command": "npx",
-      "args": ["-y", "@anthropic/mcp-server-filesystem", "/path/to/dir"]
-    },
-    "github": {
-      "transport": "stdio",
-      "command": "npx",
-      "args": ["-y", "@anthropic/mcp-server-github"],
-      "env": {
-        "GITHUB_TOKEN": "${GITHUB_TOKEN}"
-      }
-    }
-  },
-  "settings": {
-    "defaultTimeout": 30000,
-    "maxConcurrent": 5
-  }
-}
-```
-
-### MCP Commands
-
-```bash
-codesavant mcp --list              # List configured servers
-codesavant mcp add name "command"  # Add a server
-codesavant mcp remove name         # Remove a server
-```
-
-### Supported Transports
-
-- **stdio** - Standard input/output (local processes)
-- **http** - HTTP/HTTPS endpoints
-- **sse** - Server-Sent Events
-
----
-
-## Live Preview
-
-Manage development server for web projects:
-
-```
-> /preview start                    # Start dev server
-> /preview start --port 3000        # Specify port
-> /preview stop                     # Stop server
-> /preview status                   # Show status
-> /preview open                     # Open in browser
-> /preview list                     # List running servers
-```
-
-### Supported Frameworks
-
-- Next.js
-- Vite (React, Vue, Svelte)
-- Create React App
-- Angular
-- Nuxt
-- Astro
-- Gatsby
-
----
-
-## CI/CD Fix Mode
 
-Automatically fix build and test failures:
+# Plan mode (read-only)
+codesavant --plan
 
-```bash
-# Auto-fix until CI passes (max 10 attempts)
-codesavant --fix-ci
-
-# Custom max attempts
-codesavant --fix-ci --fix-ci-max-attempts 5
-
-# Specify CI commands
-codesavant --fix-ci --fix-ci-commands "npm test,npm run build"
+# Specify provider
+codesavant --provider openai --model gpt-4o
 ```
 
-CodeSavant will:
-1. Run CI commands
-2. Analyze failures
-3. Apply fixes
-4. Repeat until passing or max attempts reached
-
----
-
-## Checkpoints & Recovery
-
-### Automatic Checkpoints
-
-CodeSavant automatically creates checkpoints:
-- Before file modifications
-- Periodically during long sessions
-- On graceful shutdown
-
-### Rewind to Checkpoint
-
-Press `Esc Esc` (double escape) to rewind to the last checkpoint, restoring:
-- Conversation state
-- Modified files
-
-### Session Recovery
-
-If a session crashes, CodeSavant can recover:
-- Detect crashed sessions
-- Restore conversation state
-- Retry failed operations
-
----
-
-## Enterprise Features
-
-CodeSavant includes enterprise-grade features:
-
-| Feature | Description |
-|---------|-------------|
-| **Managed Policies** | Organization-wide tool/provider restrictions |
-| **SSO/SAML** | OAuth2 and SAML 2.0 authentication |
-| **Audit Logging** | Complete audit trail of all operations |
-| **Team Sharing** | Share sessions and workspaces |
-| **Budget Controls** | Token and cost limits per user/team |
-| **Admin Console** | Organization and user management |
-
----
-
-## Debug & Troubleshooting
-
-### Debug Logging
-
-```bash
-# Enable specific debug categories
-codesavant --debug api,mcp,tool
-
-# Enable all except MCP
-codesavant --debug all,!mcp
-
-# MCP protocol debugging
-codesavant --mcp-debug
-
-# Verbose output
-codesavant --verbose
-```
-
-### Debug Categories
-
-`api`, `mcp`, `tool`, `provider`, `stream`, `session`, `context`, `permission`, `checkpoint`, `error`, `all`
-
-### Doctor Command
-
-```bash
-codesavant doctor
-```
-
-Checks:
-- Node.js version
-- Provider configurations
-- API key validity
-- ripgrep installation
-- File permissions
-
-### Common Issues
-
-#### "No AI providers available"
-
-```bash
-codesavant config --providers
-echo $DEEPSEEK_API_KEY
-```
-
-#### npm permission errors
-
-```bash
-sudo chown -R $(whoami) ~/.npm
-npm cache clean --force
-```
+## Configuration
 
-#### Session not persisting
+Set your API key for the provider you want to use:
 
 ```bash
-ls ~/.codesavant/sessions/
-codesavant sessions -l
-```
-
----
-
-## CLI Reference
-
-```
-Usage: codesavant [options] [command] [prompt]
-
-Options:
-  -V, --version                    Output version number
-  --verbose                        Enable verbose logging
-  --debug <categories>             Enable debug logging (api,mcp,tool,...)
-  --mcp-debug                      Enable MCP protocol debugging
-  -p, --print                      Non-interactive print mode
-  -c, --continue                   Continue last session
-  -r, --resume <sessionId>         Resume specific session
-  --provider <provider>            Select provider (deepseek, groq, etc.)
-  --max-tokens <tokens>            Maximum tokens in response
-  -m, --mode <mode>                Permission mode (default, plan, strict)
-  -o, --output-format <format>     Output format (text, json, stream-json)
-  --system-prompt <prompt>         Override system prompt
-  --append-system-prompt <text>    Append to system prompt
-  --allowed-tools <tools>          Restrict to specific tools
-  --disallowed-tools <tools>       Block specific tools
-  --input-format <format>          Input format (text, stream-json)
-  --fix-ci                         CI/CD fix mode
-  --fix-ci-max-attempts <n>        Max fix attempts (default: 10)
-  --fix-ci-commands <commands>     CI commands to run
-  --max-turns <turns>              Maximum conversation turns
-  --dangerously-skip-permissions   Skip all permission checks
-  --no-mcp                         Disable MCP servers
-  --model <model>                  Specific model to use
-  --classic                        Use classic terminal UI
-  -h, --help                       Display help
-
-Commands:
-  config                           Manage configuration
-  sessions                         List and manage sessions
-  mcp                              Manage MCP servers
-  doctor                           Run diagnostic checks
-  upgrade                          Check for updates
-```
+# Anthropic (default)
+export ANTHROPIC_API_KEY=your-key
 
----
-
-## Environment Variables
-
-### API Keys
-
-| Variable | Description |
-|----------|-------------|
-| `DEEPSEEK_API_KEY` | DeepSeek API key |
-| `GROQ_API_KEY` | Groq API key |
-| `GEMINI_API_KEY` | Google Gemini API key |
-| `OPENAI_API_KEY` | OpenAI API key |
-| `ANTHROPIC_API_KEY` | Anthropic API key |
-| `GOOGLE_APPLICATION_CREDENTIALS` | Path to GCP credentials |
-
-### Shell Configuration
-
-Add to your `~/.bashrc` or `~/.zshrc`:
+# OpenAI
+export OPENAI_API_KEY=your-key
 
-```bash
-# API Keys
-export DEEPSEEK_API_KEY="your-key-here"
+# Gemini
+export GOOGLE_API_KEY=your-key
 
-# Aliases
-alias cs="codesavant"
-alias csc="codesavant -c"
+# AWS Bedrock
+export AWS_ACCESS_KEY_ID=your-key
+export AWS_SECRET_ACCESS_KEY=your-secret
+export AWS_REGION=us-east-1
 ```
 
----
+## Providers
 
-## Contributing
+| Provider | Models | Auth |
+|----------|--------|------|
+| Anthropic | claude-3-opus, claude-3-sonnet, claude-3-haiku | ANTHROPIC_API_KEY |
+| OpenAI | gpt-4o, gpt-4-turbo, gpt-3.5-turbo | OPENAI_API_KEY |
+| Gemini | gemini-1.5-pro, gemini-1.5-flash | GOOGLE_API_KEY |
+| Bedrock | claude-3, llama-3 | AWS credentials |
+| Groq | llama-3, mixtral | GROQ_API_KEY |
+| DeepSeek | deepseek-chat, deepseek-coder | DEEPSEEK_API_KEY |
+| Ollama | Any local model | Local server |
 
-Contributions are welcome! Please see our [Contributing Guide](CONTRIBUTING.md) for details.
+## Tools
 
-1. Fork the repository
-2. Create your feature branch (`git checkout -b feature/amazing-feature`)
-3. Commit your changes (`git commit -m 'Add amazing feature'`)
-4. Push to the branch (`git push origin feature/amazing-feature`)
-5. Open a Pull Request
+CodeSavant includes a comprehensive set of tools:
 
----
+- **Read**: Read file contents
+- **Write**: Create/overwrite files
+- **Edit**: Make targeted edits to files
+- **MultiEdit**: Batch edit multiple files atomically
+- **Bash**: Execute shell commands
+- **Glob**: Find files by pattern
+- **Grep**: Search file contents
+- **LS**: List directory contents
+- **LSP**: Language server integration
+- **NotebookRead**: Read Jupyter notebooks
+- **Task**: Spawn subagents for complex tasks
 
 ## License
 
-MIT License - see the [LICENSE](LICENSE) file for details.
-
----
-
-## Acknowledgments
-
-- Built with [Commander.js](https://github.com/tj/commander.js/)
-- Terminal UI powered by [Ink](https://github.com/vadimdemedes/ink) and [Chalk](https://github.com/chalk/chalk)
-- AI integrations via [OpenAI SDK](https://github.com/openai/openai-node)
-- Fast search with [ripgrep](https://github.com/BurntSushi/ripgrep)
-
----
-
-<div align="center">
-
-**Made with love by [Vishal Aggarwal](https://github.com/burnisback)**
-
-[Report Bug](https://github.com/burnisback/codesavant/issues) · [Request Feature](https://github.com/burnisback/codesavant/issues)
-
-</div>
+MIT