GitHub radar
FrontierAgent — A multi-agent research framework
Terminal TUI for an open-source multi-agent system. A coordinator distributes work to parallel sub-agents, aggregates their responses, and composes the final output.
Apodex AI’s FrontierAgent is an open-source multi-agent runtime and terminal TUI designed for long-horizon tasks such as research, file analysis, and document writing. In addition to the one-shot ReAct workflow mode (where a single stateful agent performs the entire task), there’s also the Agent Team mode, which lets you break down the task into parallel subtasks and delegate them to sub-agents, then aggregate their outputs and compose the final result. You can see real-time progress of individual tasks, currently running tools, and past actions. Mutative actions will prompt for diff approval and allow you to save checkpoints along the way, making your session resumable. All file I/O is sandboxed so that input files are immutable, your workspace is at /workspace, and all persistent outputs are placed at /outputs. FrontierAgent supports connecting to any OpenAI compatible API server (e.g., Apodex, OpenAI, or your own local model server). Open sourced under Apache 2.0.
Why a vibe-coder should care
When you need to delegate a complex, multi-step task to an agent and move on, FrontierAgent has all the features you need for isolation, parallelization, and approving high-risk actions. Use it to automate your research, code audits, or long document creation that requires more than a single shot conversation.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Install FrontierAgent from https://github.com/ApodexAI/FrontierAgent — follow the README (requires Python 3.12 and uv), create a .env with your OpenAI-compatible API key, launch in agent_team mode, and show me the TUI. Ask me for the key if you need one.
Requires Python 3.12 and uv. With an external API key (OpenAI, Apodex, or another compatible provider) any laptop will do. To run a model locally, you'll need a powerful NVIDIA GPU.
Open on GitHub▌ More finds