agentrove
Self-hosted AI coding workspace for running and orchestrating Antigravity, Claude Code, Codex, Copilot, Cursor, Grok, and OpenCode agents from one interface.
Ficha actualizada 22 sept 2026, 18:47 · upstream 22 sept 2026, 18:02 · release v0.1.60
Qué es agentrove
Descripción editorial construida a partir del repositorio, su documentación y sus señales de mantenimiento.
Agentrove is a self-hosted AI coding workspace that runs and orchestrates multiple AI agents including Antigravity, Claude Code, Codex, Copilot, Cursor, Grok, and OpenCode. It provides isolated Docker or host sandboxes for each agent, combining chat, code editor, terminal, file tree, diffs, secrets management, and git tools into a unified workspace.
The platform supports workspaces from empty folders to existing Git repositories on GitHub. Agents can stream sessions with cancellation support, permission prompts, queued follow-up messages, file mentions, slash commands, and attachments. Sub-threads allow parallel execution of tasks within a lead chat, while per-turn model and persona selection enables flexible workflow patterns.
The bundled MCP server exposes the entire instance as tools, enabling any chat to act as an orchestrator that decomposes work into sub-tasks routed to appropriate agents. Worktrees are isolated via git for concurrent editing without conflicts. Polling mechanisms handle turn completion and rework distribution between workers.
- Run multiple AI coding assistants simultaneously with isolated sandboxes
- Create multi-agent workflows where a lead agent orchestrates sub-tasks across different models and personas
- Build a unified code review environment combining chat, diffs, secrets management, and git operations
- Deploy custom agent fleets using the same orchestration framework without platform-specific configuration
- Run AI agents in isolated environments to prevent interference between concurrent executions
- Supports multiple AI coding models including Antigravity, Claude Code, Codex, Copilot, Cursor, Grok, and OpenCode through ACP adapters
- Provides isolated Docker or host sandboxes for each agent to prevent workspace interference
- Bundled MCP server allows any chat to act as an orchestrator with tools like send_message, get_messages, list_models, and list_personas
- Supports sub-threads, per-turn model/persona selection, worktree isolation, polling-based turn management, and unattended turns
- Includes GitHub-assisted repository browsing, pull request review, PR creation, reviewer selection, and branch/commit/pull helpers
- Available as Docker web app, macOS desktop app with Tauri sidecar, and native iOS app built with Tauri
- Requires Docker or a host machine to run the backend services; no true cloud-native deployment option is provided
- GitHub-assisted repository browsing requires an active GitHub account and may have rate limits
- The macOS desktop app requires Xcode installation on Mac systems, which can be a barrier for non-Mac users
- iOS app requires Xcode development tools and CocoaPods; building from source is complex for end users
- No built-in multi-language code support beyond the supported languages of each individual agent model
Perfil de infraestructura
Estimación inicial derivada del repositorio y de los servicios detectados.
Despliegue detectado
Qué señales reales de instalación hemos encontrado.
For Docker deployment, use a production-ready compose file with environment variables for SECRET_KEY, APP_URL, ALLOWED_ORIGINS, and database configuration. The backend uses FastAPI with SQLAlchemy and Redis for caching. For the macOS desktop app, ensure Xcode is installed on Mac systems to build the Tauri sidecar. The iOS app requires signing certificates from Apple Developer Program; use the APPLE_DEVELOPMENT_TEAM environment variable in the frontend config. Always set a strong SECRET_KEY using openssl or an external key management system.
Stack y dependencias
Componentes inferidos a partir de manifests, configuración y documentación.
Puede sustituir a
Relaciones detectadas por el análisis del repositorio y su categoría.
Experiencia real de la comunidad
Ayuda a contrastar la estimación automática con instalaciones reales.
Todavía no hay datos comunitarios para este proyecto.