git clone https://github.com/hem9984/Azlon-demo.git
cd Azlon-demo
echo "OPENAI_KEY=sk-..." > .env
docker compose up
- Frontend UI: http://localhost:8080/
- Restack UI: http://localhost:5233/
- Enter your user_prompt and test_conditions.
- Click "Run Workflow".
- Wait for your project code to complete!
- 🤖 It will recursively generate code, run the code, and fix the code if needed until it deems that your test case(s) are fulfilled.
This project sets up an autonomous coding workflow using Restack, OpenAI’s GPT models, Docker-in-Docker for building and running Docker images, and a frontend React UI to interact with the system. Users can provide a user_prompt and test_conditions to generate code automatically, run it in a containerized environment, and validate the results. Users can also toggle an "advanced mode" to edit system prompts directly.
- Ensure Docker (>= 20.10) and Docker Compose (>= 1.29) are installed.
- Install Docker | Install Docker Compose
- Sign up for OpenAI and get an API key: OpenAI API Keys
- Clone the Repository: bash
git clone https://github.com/hem9984/Azlon-demo.git
cd Azlon-demo
- Environment Variables: Create a .env file in the project root. Add your OpenAI key: bash
echo "OPENAI_KEY=sk-..." > .env
- Ensure this .env file contains OPENAI_KEY.
- restack-engine
- docker-dind (Docker-in-Docker)
- backend (FastAPI + Restack)
- frontend (React UI)
- Start them all with one command: bash
docker compose up
- Run Restack engine on http://localhost:5233 (and other ports as specified).
- Run the backend on http://localhost:8000
- Run the frontend on http://localhost:8080
- Check the docker-compose.yml and frontend/Dockerfile for the final frontend port and mode.
Restack UI: http://localhost:5233/
The Restack UI will show you running workflows and other details.
Frontend UI: http://localhost:8080/
The React UI lets you enter your user_prompt and test_conditions. If you enable advanced mode in the GUI, you can edit system prompts as well.
- Open the frontend URL (http://localhost:8080/) in your browser. a. Also open http://localhost:5233/ if you want to view the process in real-time!
- Enter your user_prompt and test_conditions. a. Toggle advanced mode if you want to modify the system prompts.
- Click "Run Workflow".
- Use your prompts to generate code and a Docker environment.
- Build and run the generated code in a container.
- Validate the output against test_conditions.
- Display the results in the frontend UI.