PilotDeck
Active·★ 2.0k·AGPL-3.0·Updated 2026-05-29
★ Trending★ Multi-Agent
Task-oriented AI Agent productivity platform
PilotDeck is an open-source agent operating system built around the WorkSpace concept, jointly developed by Tsinghua University THUNLP, ModelBest, OpenBMB, and AI9Stars. It targets long-running, multi-project productivity work with three core capabilities: white-box traceable memory, smart model routing for cost optimization, and always-on background task execution. The system natively supports MCP and works consistently across Web, CLI, and IM frontends.
#multi-agent
01
Features
01WorkSpace-level isolation: each project has its own file system, memory store, and skill set
02White-box memory with end-to-end visibility, Dream Mode consolidation, and one-click rollback
03Smart routing auto-detects task difficulty and dispatches to appropriate models to reduce token cost
04Always-on background execution: agent continues discovering and running tasks after user steps away
05Native MCP support, consistent behavior across Web, CLI, and IM frontends
02
Compatibility
macOS
Native
Verified via docs
Linux
Native
Verified via docs
Docker
Compose
Verified via docs
MCP Protocol
Native
Verified via docs
Claude / GPT-4o
Smart Routing
Verified via docs
03
Quick start
1
$ curl -fsSL https://raw.githubusercontent.com/OpenBMB/PilotDeck/main/install.sh | bash
04
Use cases
↳Long-running multi-project knowledge work with isolated memory and skill sets per project
↳Cost-optimized AI agent pipelines that automatically route tasks to models based on complexity
↳Background autonomous execution of monitoring, research, or batch tasks that deliver results as local files
05
Alternatives
Related searches
Comments
Log in to leave a comment
No comments yet. Be the first!