devorch 0.2.1

Multi-provider AI coding assistant CLI with 13+ providers

DevOrch

A multi-provider AI coding assistant CLI, similar to Claude Code and Gemini CLI.

Screenshots

Startup	Chat

Provider Selection	Model Selection

Tool Execution	Terminal Session

Features

• 13 AI Providers - OpenAI, Anthropic, Gemini, Groq, Mistral, Together AI, OpenRouter, GitHub Copilot, DeepSeek, Kimi, Ollama, LM Studio, and Custom
• Memory System - Persistent memory across conversations (user preferences, feedback, project context)
• MCP Support - Connect Model Context Protocol servers for extensible tools
• Skills System - Built-in and custom skills (/commit, /review, /test, /fix, /explain, /simplify)
• Terminal Sessions - Background processes with GUI option, persist across restarts
• Dynamic Model Listing - Fetches latest available models from provider APIs
• Secure API Key Storage - Uses system keychain (Windows, macOS, Linux)
• Session Persistence - SQLite-based chat history with resume capability
• Powerful Tools - Shell, terminal sessions, file operations, search, grep, code editing, web access, memory
• Interactive UI - Arrow-key navigation, numbered selections, bottom status bar, markdown responses
• Permission System - Configurable allow/deny rules with interactive prompts
• In-Chat Auth - Set or update API keys with /auth without restarting
• Multiple Modes - Plan mode, Auto mode, and Ask mode

Installation

Option 1: Install with pipx (Recommended)

pip install pipx
pipx install devorch

Option 2: Install with pip

pip install devorch

Option 3: Install from source

git clone https://github.com/Amanbig/DevOrch.git
cd DevOrch
pip install -e .

Quick Start

# Start DevOrch (first run shows interactive setup)
devorch

# Specify a provider
devorch -p openai
devorch -p anthropic
devorch -p local  # Ollama

Slash Commands

Navigation & Switching

Command	Description
/help	Show all commands grouped by category
/models	Browse and switch models (interactive, numbered list)
/model <name>	Switch model (supports partial match, e.g. /model opus)
/providers	Browse and switch providers (interactive, numbered list)
/provider <name>	Switch provider directly
/mode	Switch between Plan/Auto/Ask modes
/status	Show provider, model, mode, memories, skills, MCP
/auth [provider]	Set or update API key for current or specified provider

Memory

Command	Description
/memory	Show all saved memories
/remember <text>	Save something to memory
/forget	Delete a memory (interactive)

Memories persist across conversations in ~/.devorch/memory/. Types:

• user - Your role, preferences, expertise
• feedback - Corrections and guidance for the AI
• project - Project decisions, context, ongoing work
• reference - Links to external resources

Skills

Command	Description
/skills	List all available skills
/skill <name>	Run a skill
/commit	Create a git commit with a descriptive message
/review	Review code changes for bugs and issues
/test	Run project tests and analyze results
/fix	Fix the last error or failing test
/explain	Explain the current project structure
/simplify	Simplify recent code changes

Custom skills can be added as YAML files in ~/.devorch/skills/:

# ~/.devorch/skills/deploy.yaml
name: deploy
description: Deploy to production
prompt: |
  Run the deploy script and verify it succeeds.
  Check the deploy logs for any errors.

Session

Command	Description
/session	Show current session info
/history	Show conversation history
/clear	Clear conversation history
/compact	Summarize and compact history
/save	Save conversation to file
/undo	Undo last message

MCP

Command	Description
/mcp	Show connected MCP servers and their tools

MCP (Model Context Protocol)

Connect external tool servers via MCP. Configure in ~/.devorch/config.yaml:

mcp_servers:
  filesystem:
    command: npx
    args: ["-y", "@modelcontextprotocol/server-filesystem", "/home/user"]
  github:
    command: npx
    args: ["-y", "@modelcontextprotocol/server-github"]
    env:
      GITHUB_TOKEN: "ghp_xxx"
  sqlite:
    command: uvx
    args: ["mcp-server-sqlite", "--db-path", "mydb.sqlite"]

MCP tools automatically appear alongside built-in tools and can be used by the AI.

Modes

• ASK (default) - Asks before each tool execution
• AUTO - Executes tools automatically (dangerous commands still blocked)
• PLAN - Shows plan before executing, asks for approval

Terminal Sessions

Unified terminal tool with optional GUI window. The AI can monitor output and send input.

> Start a dev server (headless, AI monitors)
  > terminal_session start command="npm run dev"
    ✓ Session 'swift-fox-a3f2' started (PID 12345)

> Open a visible terminal (user can type, AI reads output)
  > terminal_session start command="bash" gui=true
    ✓ Session 'calm-owl-b7e1' started in visible terminal

> Check output
  > terminal_session read session_id="swift-fox-a3f2"
    ✓ [Session 'swift-fox-a3f2' — running]

> Stop it
  > terminal_session stop session_id="swift-fox-a3f2"
    ✓ Session stopped.

Sessions are tracked in ~/.devorch/sessions/ and can be reconnected after restarting DevOrch.

Tool Permissions

Interactive permission system with arrow-key navigation:

? Choose an action:
  » Allow once
    Allow for this session
    Always allow (save to config)
    Deny

devorch permissions list                    # Show permissions
devorch permissions set shell allow         # Always allow shell
devorch permissions allow shell "git *"     # Allow git commands
devorch permissions deny shell "rm -rf *"   # Block dangerous commands

Supported Providers

Cloud Providers

Provider	Models	API Key
OpenAI	GPT-4o, GPT-4, o1	OPENAI_API_KEY
Anthropic	Claude 4, Claude 3.5	ANTHROPIC_API_KEY
Google Gemini	Gemini 2.0, 1.5 Pro/Flash	GOOGLE_API_KEY
Groq	Llama 3.3, Mixtral	GROQ_API_KEY
Mistral	Large, Codestral	MISTRAL_API_KEY
Together AI	Llama 3, Mixtral, Qwen	TOGETHER_API_KEY
OpenRouter	100+ models	OPENROUTER_API_KEY
GitHub Copilot	GPT-4o, Claude 3.5	GITHUB_TOKEN
DeepSeek	Chat, Coder, Reasoner	DEEPSEEK_API_KEY
Kimi (Moonshot)	8K, 32K, 128K context	MOONSHOT_API_KEY

Local & Self-Hosted

Provider	Setup
Ollama	Install Ollama, run models locally
LM Studio	Install LM Studio, GUI for local models
Custom	Any OpenAI-compatible API (vLLM, TGI, llama.cpp)

Tools

Tool	Description
shell	Execute shell commands
terminal_session	Managed sessions with optional GUI (start, read, send, stop)
filesystem	Read, write, and list files
search	Find files by name patterns (glob)
grep	Search for text patterns in files
edit	Make targeted edits to existing files
task	Track progress on multi-step work
memory	Save/search persistent memories across conversations
websearch	Search the web (DuckDuckGo)
webfetch	Fetch content from a URL

Configuration

API Keys

# From CLI
devorch set-key openai
devorch set-key anthropic

# From within chat
/auth openai
/auth anthropic

# Or use environment variables
export OPENAI_API_KEY=sk-...
export ANTHROPIC_API_KEY=sk-ant-...
export GOOGLE_API_KEY=...
export GROQ_API_KEY=gsk_...

Config File

~/.devorch/config.yaml:

default_provider: openai

providers:
  openai:
    default_model: gpt-4o
  anthropic:
    default_model: claude-sonnet-4-20250514

# MCP servers
mcp_servers:
  github:
    command: npx
    args: ["-y", "@modelcontextprotocol/server-github"]
    env:
      GITHUB_TOKEN: "ghp_xxx"

Directory Structure

~/.devorch/
├── config.yaml          # Provider settings, MCP servers
├── permissions.yaml     # Tool permission rules
├── sessions.db          # Chat history (SQLite)
├── memory/              # Persistent memories
│   ├── MEMORY.md        # Memory index
│   ├── user_*.md        # User profile memories
│   ├── feedback_*.md    # Feedback memories
│   └── project_*.md     # Project context memories
├── skills/              # Custom skill definitions
│   └── *.yaml           # User-defined skills
└── sessions/            # Terminal session logs
    ├── registry.json    # Session registry
    └── *.log            # Session output logs

Requirements

• Python 3.10+
• Dependencies: typer, rich, pydantic, openai, anthropic, google-genai, httpx, keyring, prompt_toolkit, questionary, pyyaml, duckduckgo-search

License

MIT