Summary

OpenDev is an open-source, terminal-native coding agent architected as a compound AI system. It addresses the need for flexible LLM integration by enabling users to orchestrate specialized sub-agents, each bound to a distinct LLM, for tasks like execution, reasoning, and summarization. This allows fine-grained control over model capabilities, latency, and cost, empowering developers with a proactive, autonomous coding assistant.

How It Works

OpenDev utilizes a structured ensemble of agents and workflows, moving beyond monolithic LLM architectures. Each workflow (Execution, Thinking, Compaction, Self-Critique, VLM) can be independently bound to a user-selected LLM from any supported provider. This compound AI approach optimizes resource allocation, e.g., using a powerful model for reasoning and a cheaper one for summarization. Written in Rust, it prioritizes high performance and minimal resource consumption with a small, standalone binary, organizing work into concurrent sessions for parallel agent execution.

Quick Start & Requirements

Installation requires cloning the repository, building from source (cargo build --release -p opendev-cli), and copying the binary. Users must set API keys for supported LLM providers. Launch the TUI with opendev or the Web UI with opendev run ui. Interactive setup is available via opendev config setup. A Provider Setup Guide is referenced for detailed configuration.

Highlighted Details

• Multi-Provider Orchestration: Supports 9 LLM providers (OpenAI, Anthropic, Google, Mistral, etc.).
• Modular Workflows: Five workflow types (Normal, Thinking, Compact, Critique, VLM) can be assigned unique LLMs for tailored performance and cost.
• Model Context Protocol (MCP): Enables dynamic discovery and integration of external tools.
• Performance: Rust implementation yields a ~3.7MB binary with instant startup and low memory usage.
• Dual Interface: Offers both a Terminal UI (TUI) and a Web UI for remote monitoring.

Maintenance & Community

Contributions are welcomed via issues and pull requests. However, specific community channels and maintainer details are absent from the README. Official website and comprehensive documentation are noted as "coming soon."

Licensing & Compatibility

The README omits explicit license information, creating ambiguity regarding usage rights, modification, and distribution, particularly for commercial applications.

Limitations & Caveats

The project is under active development, with pending website and documentation. The lack of a stated license is a significant adoption blocker. Users must build from source and manage LLM API key configurations. The original Python version is archived.