dora-crew Documentation

Rust-Native Coding Agent Orchestration Framework

"Ship code with your crew"

---

Documents

Document	Description
01-insight-report.md	Industry analysis: vibe coding, Gastown, agent orchestration trends, and counter-arguments
02-prd.md	Product Requirements Document: goals, user stories, functional requirements
03-architecture.md	Technical architecture: components, data flow, crate structure, integrations
04-roadmap.md	Implementation roadmap: phases, milestones, tasks, resource requirements

---

Quick Links

Key Concepts

• dora-crew: A Rust-native coding agent orchestration framework built on dora-rs
• Supervisor-Worker Pattern: Simple 2-role hierarchy for multi-agent coordination
• Context Engineering: Structuring information so AI agents produce optimal outputs
• Memory System: Working + episodic + semantic memory for persistent agent knowledge

Timeline

Phase	Duration	Goal
Phase 0	Weeks 1-4	Foundation & spikes
Phase 1	Weeks 5-12	Single agent works
Phase 2	Weeks 13-20	Multi-agent works
Phase 3	Weeks 21-28	Memory & MCP
Phase 4	Weeks 29-36	v1.0 release

Key Decisions

Decision	Choice
Foundation	dora-rs
Language	Rust (100%, no C deps)
LLM abstraction	Rig-compatible trait
Task storage	native_db (pure Rust)
Episodic memory	DataFusion + Parquet
Semantic memory	Lance (vector search)
Protocol	MCP (Model Context Protocol)

---

Getting Started

# Clone the repository
git clone https://github.com/dora-rs/dora-crew

# Read the documents in order
1. Start with the Insight Report for context
2. Review the PRD for requirements
3. Study the Architecture for technical design
4. Follow the Roadmap for implementation

---

Contributing

See CONTRIBUTING.md for guidelines.

---

Last Updated: January 2026