Skip to content

QUICKSTART.md

Latest commit

 

History

History
306 lines (233 loc) · 6.85 KB

QUICKSTART.md

File metadata and controls

306 lines (233 loc) · 6.85 KB

🚀 Quick Start Guide - AutoCodeAI v2.0

Get up and running with AutoCodeAI in 5 minutes!

Prerequisites

  • Docker (for sandboxed code execution)
  • Python 3.11+
  • API Key for your chosen LLM provider (OpenAI, DeepSeek, or local Ollama)

Installation

Option 1: Web UI (Recommended for First-Time Users)

# 1. Clone the repository
git clone https://github.com/yourusername/autocodeai.git
cd autocodeai

# 2. Install dependencies
pip install -r requirements.txt

# 3. Configure environment
cp .env.example .env
# Edit .env and add your API key:
# For OpenAI:
#   LLM_MODE=openai
#   OPENAI_API_KEY=sk-your-key-here
# For DeepSeek:
#   LLM_MODE=deepseek
#   DEEPSEEK_API_KEY=your-key-here

# 4. Start the server
uvicorn main:app --reload

# 5. Open the web UI
open http://localhost:8000

You should see the AutoCodeAI web interface! 🎉

Option 2: Docker Compose (Full Stack)

# 1. Clone and configure
git clone https://github.com/yourusername/autocodeai.git
cd autocodeai
cp .env.example .env
# Edit .env with your API keys

# 2. Start everything
docker compose up --build

# 3. Access the UI
open http://localhost:8000

This starts:

  • AutoCodeAI backend (port 8000)
  • ChromaDB vector database (port 8001)
  • Web UI

Your First Task

Using the Web UI

  1. Open http://localhost:8000
  2. Type a task in the text area: 
    Create a REST API endpoint for user registration with email validation,
password hashing, and proper error handling. Include unit tests.
  3. Click "🚀 Run Agents"
  4. Watch the agents work in real-time!

Using the API (curl)

curl -X POST http://localhost:8000/api/agent/run \
  -H "Content-Type: application/json" \
  -d '{
    "task": "Write a binary search function with comprehensive pytest tests",
    "context_files": []
  }'

Using Python

import asyncio
from services.orchestrator import Orchestrator

async def main():
    orch = Orchestrator()
    
    results = await orch.run(
        task="Create a FastAPI endpoint for file upload with size validation",
        context_files=[],
        callback=lambda msg: print(msg, end="", flush=True)
    )
    
    print("\n\n✅ Task complete!")
    for result in results:
        print(f"- {result['type']}: {result['step']}")

asyncio.run(main())

Example Tasks

Try these examples to see AutoCodeAI in action:

1. Code Generation

Create a Python class for managing a TODO list with add, remove, complete, 
and list methods. Include full pytest coverage with fixtures.

2. API Development

Build a FastAPI endpoint for user authentication using JWT tokens. 
Include login, register, and protected routes. Add unit tests.

3. Data Processing

Write a function to parse CSV files and convert them to JSON with 
error handling for malformed data. Include edge case tests.

4. Bug Fixing

File: buggy_sort.py
Task: Fix the sorting algorithm and add comprehensive tests

Advanced Features

🔧 Tool Integration

Create tasks that use git, pip, and shell commands:

{
  "task": "Clone the FastAPI repo and create a hello world example",
  "context_files": []
}

The planner can include tool steps:

{
  "steps": [
    {
      "agent": "tool",
      "tool_name": "git_clone",
      "tool_params": {
        "url": "https://github.com/tiangolo/fastapi",
        "dest": "fastapi_lib"
      }
    },
    {
      "agent": "coder",
      "description": "Create hello world using FastAPI"
    }
  ]
}

⚡ Parallel Execution

For tasks with independent subtasks, the planner automatically groups them:

Create three microservices: user service, auth service, and notification service.
Each should have its own API endpoint and tests.

The planner creates parallel groups:

{
  "steps": [
    {"agent": "coder", "description": "User service", "parallel_group": 1},
    {"agent": "coder", "description": "Auth service", "parallel_group": 1},
    {"agent": "coder", "description": "Notification service", "parallel_group": 1},
    {"agent": "tester", "description": "Integration tests", "parallel_group": 2}
  ]
}

All three services are coded in parallel, then tested together!

🌐 Switching LLM Providers

Use DeepSeek (Faster, Cheaper for Coding)

# Edit .env
LLM_MODE=deepseek
DEEPSEEK_API_KEY=sk-your-deepseek-key

# Or set environment variables
export LLM_MODE=deepseek
export DEEPSEEK_API_KEY=sk-...

Use Local Models (Free, Private)

# 1. Install and start Ollama
# Download from https://ollama.ai

# 2. Pull a model
ollama pull deepseek-coder

# 3. Configure AutoCodeAI
export LLM_MODE=local
export LOCAL_MODEL=deepseek-coder
export LOCAL_LLM_URL=http://localhost:11434/v1

# 4. Run!
uvicorn main:app --reload

Mix and Match Models

Use different models for different agents:

# .env
LLM_MODE=litellm
PLANNER_MODEL=gpt-4o              # Best for planning
CODER_MODEL=deepseek/deepseek-chat # Fast for coding
TESTER_MODEL=groq/llama-3.3-70b-versatile  # Free, fast
DEBUGGER_MODEL=anthropic/claude-sonnet-4-5 # Best for debugging
CRITIC_MODEL=anthropic/claude-sonnet-4-5   # Best for review

Configuration Tips

Minimum Config (OpenAI)

LLM_MODE=openai
OPENAI_API_KEY=sk-your-key-here

Cost-Optimized Config

LLM_MODE=litellm
PLANNER_MODEL=gpt-4o-mini           # Cheaper planner
CODER_MODEL=deepseek/deepseek-chat  # Very cheap coding
TESTER_MODEL=groq/llama-3.3-70b-versatile  # Free!
DEBUGGER_MODEL=deepseek/deepseek-chat
CRITIC_MODEL=gpt-4o-mini

Privacy-First Config (All Local)

LLM_MODE=local
LOCAL_LLM_URL=http://localhost:11434/v1
LOCAL_MODEL=deepseek-coder
ENABLE_PARALLEL_EXECUTION=true

Troubleshooting

"API key not found"

  • Check your .env file exists
  • Verify the API key variable matches your LLM_MODE
  • Restart the server after changing .env

"Docker not found"

  • Install Docker Desktop
  • Ensure Docker daemon is running
  • Test: docker ps

Slow response

  • Try using DeepSeek (LLM_MODE=deepseek) - it's faster
  • Enable parallel execution in .env
  • Use local models for maximum speed

Web UI not loading

  • Check the server is running: curl http://localhost:8000/health
  • Look for errors in the terminal
  • Try clearing browser cache

Next Steps

  1. Read the full README.md for architecture details
  2. Check CHANGELOG.md for all v2.0 features
  3. Explore the API at http://localhost:8000/docs
  4. Join the community (link to Discord/GitHub Discussions)
  5. Contribute - see CONTRIBUTING.md

Getting Help

Happy Coding! 🚀 If AutoCodeAI helps you, please star the repo ⭐