Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
🤖 AI-powered software engineering multi-agent system with researcher and developer agents that automate code implementation through intelligent planning and execution. Built with LangGraph multi-agent workflows
| Date | Stars |
|---|---|
| 2026-07-31 | 637 |
| 2026-08-01 | 637 |
| 2026-08-06 | 637 |
| 2026-08-12 | 638 |
| 2026-08-18 | 638 |
| 2026-08-20 | 638 |
| 2026-08-21 | 638 |
| 2026-08-24 | 638 |
| 2026-08-28 | 638 |
| 2026-09-07 | 639 |
| 2026-09-09 | 639 |
| 2026-09-10 | 639 |
| 2026-09-14 | 640 |
| 2026-09-16 | 641 |
| 2026-09-20 | 641 |
Today
— stars today
This week
+2 stars this week
This month
+3 stars this month
Momentum
0.0
growth rate 0.31%/day
# SWE Agent with LangGraph - by [LangTalks](https://langtalks.ai)
<img src="./static/cover.png" width="400" alt="Cover">
  
A sophisticated AI-powered software engineering agent that automates code implementation through intelligent planning and execution. Built with LangGraph for reliable multi-agent workflows.
> ⚠️ **Alpha Status**: This project is in active development. Features may change and some functionality is experimental. Perfect for early adopters and contributors who want to shape the future of AI-powered development.
[end to end showcase](https://youtu.be/vJNqAgLzOSg)
## 🚀 Features
- **Intelligent Code Planning**: AI architect analyzes requirements and creates detailed implementation plans
- **Automated Code Generation**: Developer agent executes plans with precise file modifications
- **Multi-Agent Workflow**: Separate planning and implementation phases for better reliability
- **Codebase Understanding**: Advanced code analysis using tree-sitter and semantic search
- **Incremental Development**: Atomic task breakdown for safer, more manageable changes
## 🏗️ Architecture
The system uses a two-stage LangGraph workflow:
### 1. Architect Agent - Research & Planning
The architect agent:
- Researches the codebase structure and patterns
- Analyzes requirements and creates hypotheses
- Generates detailed implementation plans with atomic tasks
- Uses tools for code search and semantic understanding
### 2. Developer Agent - Implementation
The developer agent:
- Executes implementation plans step by step
- Performs atomic code modifications with precision
- Creates new files and modifies existing ones
- Validates changes against the original requirements
### Workflow Overview
```
User Request → Architect (Research & Plan) → Developer (Implement) → Results
```
**Key Components:**
- **State Management**: Structured data flow between agents using Pydantic models
- **Tool Integration**: File system operations, code search, and structure analysis
- **Research Pipeline**: Hypothesis-driven exploration of codebases
- **Atomic Execution**: Granular tasks for reliable implementation
## 🔄 Agent State Management
The system uses a hierarchical state management approach with Pydantic models for type safety and validation. Each agent maintains its own state while sharing common entities for seamless data flow.
### Main Agent State (`AgentState`)
The top-level orchestrator state that manages the overall workflow:
```python
class AgentState(BaseModel):
implementation_research_scratchpad: Annotated[list[AnyMessage], add_messages]
implementation_plan: Optional[ImplementationPlan] = None
```
**Fields:**
- `implementation_research_scratchpad`: Message history from research and planning phase
- `implementation_plan`: Structured plan created by architect agent for developer execution
### Architect Agent State (`SoftwareArchitectState`)
Manages the research and planning phase with hypothesis-driven exploration:
```python
class SoftwareArchitectState(BaseModel):
research_next_step: Optional[str] = None
implementation_plan: Optional[ImplementationPlan] = None
implementation_research_scratchpad: Annotated[list[AnyMessage], add_messages] = []
is_valid_research_step: Optional[bool] = None
```
**Fields:**
- `research_next_step`: Current hypothesis or research direction being explored
- `implementation_plan`: Generated structured plan with atomic tasks
- `implementation_research_scratchpad`: Research conversation history and tool outputs
- `is_valid_research_step`: Validation flag for research hypothesis quality
**Workflow:**
1. Generate research hypothesis → Validate hypothesis → Conduct research → ExtrExcerpt of 15,798 characters
Read on GitHub
Lee Twito
9
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:26b4f8e48ba06b44, desc:multi-agent, desc:multi agent, desc:planning and execution