traycerai/traycer

Traycer – AI‑Orchestrated Coding Assistant

What it is – Traycer is a desktop application that lets you run multiple LLM‑based coding agents side‑by‑side. It works as a thin orchestration layer: you keep the subscriptions you already have (OpenAI, Anthropic, etc.) and Traycer routes requests to those providers while sharing a single “context window” across them. The UI adds collaboration tools (shared boards, ticket assignment) and syncs the agent state across devices.

Key capabilities

• Bring‑Your‑Own‑Agent (BYOA) – Connect any of the listed coding agents (Claude Code, OpenAI Codex, GitHub Copilot, etc.) or use Traycer’s own inference service.
• Unified context – Switch models inside the same agent without losing the conversation history; the memory is shared across providers.
• Agent‑to‑Agent communication – Agents can talk to each other automatically, enabling loops such as code review or architectural debate.
• Team collaboration – Share a workspace, edit code together in real time, assign tickets to agents, and view progress on boards.
• Cross‑device sync – The same agent state is available on macOS, Linux and Windows.

How you get it – Pre‑built binaries are provided for macOS (Apple Silicon & Intel), Linux (AppImage, .deb, .rpm) and Windows (x64). The README links directly to the latest release assets.

Privacy – Code is processed only in memory; prompts are not stored unless the user disables the default Privacy Mode. Provider calls go straight to the chosen API; Traycer’s own inference runs on its own service.

Who it’s for – Developers who want to experiment with multiple coding LLMs without juggling separate tools, and teams that need a shared, real‑time AI‑assisted coding environment.

Documentation & community – Full docs, a Discord server, Twitter, and YouTube channel are linked from the README. Contributions are welcomed under an MIT license.