2.3 KiB
2.3 KiB
Community ADE (Agentic Development Environment)
A self-hosted, SDK-powered orchestration platform for Letta stateful agents. Built and directed by Ani (Annie Tunturi).
What This Is
A web dashboard + orchestration engine for managing Letta agents — their configuration, memory, tools, conversations, and background task execution. Everything flows through the official Letta SDKs.
Quick Start
cd community-ade-wt/mvp-unified
docker compose up --build -d
# Dashboard: http://10.10.20.19:4422/
# API: http://10.10.20.19:4421/api/agents
New to this project? Read START_HERE.md in the repo root first.
Architecture
- React dashboard — agent grid, config editor, chat, models, settings
- Express backend — REST API proxying to Letta via
@letta-ai/letta-client, interactive sessions via@letta-ai/letta-code-sdk - Redis — persistent task queue for background agent execution
- Letta server — the stateful agent runtime (v0.16.6)
SDK-First Design
Every feature uses one of two packages:
@letta-ai/letta-client— REST CRUD (agents, memory, tools, models)@letta-ai/letta-code-sdk— interactive sessions (chat, streaming, tool execution)
No standalone implementations. No mock data. See START_HERE.md for the full SDK compliance protocol.
Differentiation
Unlike commercial alternatives:
- Open source and self-hostable
- Stateful by design — agents remember across sessions via Letta's hierarchical memory
- Model agnostic — any OpenAI-compatible provider
- Orchestration-ready — Redis task queue with worker pool for background agent work
Documentation
| Document | What |
|---|---|
START_HERE.md |
Project orientation, SDK rules, build sequence |
docs/ade-research.md |
Competitive analysis (Letta vs Intent vs Warp) |
docs/ade-phase1-orchestration-design.md |
Task queue architecture |
docs/ade-phase1-execution-plan.md |
6-week execution plan |
docs/community-ade-research-synthesis-2026-03-18.md |
Technical patterns + gap analysis |
docs/community-ade-project-state.md |
Current project state and build progress |
License
MIT
Project orchestrated by Ani. Research by specialized subagents. Implementation driven by SDK-first architecture.