OnyxMunk/autogpt-space

AutoGPT Space

An autonomous AI agent powered by AutoGPT and Google Gemini, deployed as a Hugging Face Space using Docker.

Features

• —🤖 Autonomous Task Execution - Break down complex tasks into steps and execute them autonomously
• —🔄 Iterative Problem-Solving - Multi-step reasoning and execution powered by Google Gemini
• —🎛️ Configurable Parameters - Adjust iterations, temperature, and more
• —🌐 Beautiful Web Interface - Modern Gradio UI with progress tracking
• —🔒 Secure API Key Management - Environment variables for sensitive credentials
• —🚀 Powered by Google Gemini - Leveraging Google's advanced language models

Usage

1. 1.Configure API Keys (Required for full functionality):
2. 2.Get your Google Gemini API key from Google AI Studio
3. 3.Go to Space Settings → Repository secrets
4. 4.Add GOOGLE_API_KEY with your Google Gemini API key
5. 5.Optionally set GEMINI_MODEL (default: gemini-pro)
6. 6.Available models: gemini-pro, gemini-pro-vision

1. 1.Run Tasks:
2. 2.Enter a task description in the text box
3. 3.Adjust max iterations and temperature if needed
4. 4.Click "Run AutoGPT" to start execution
5. 5.View real-time progress and results

Configuration

Environment Variables

Add these in Space Settings → Repository secrets:

\* Required for full functionality. The Space will run in demo mode without API keys.

Get your API key: Visit Google AI Studio to create a free API key.

Available Models

• —gemini-pro - Standard text generation model (default)
• —gemini-pro-vision - Multimodal model with vision capabilities

Parameters

• —Max Iterations: Controls how many steps AutoGPT will take (1-20)
• —Temperature: Controls the randomness of LLM responses (0.0-2.0)
• —Lower values (0.0-0.5): More deterministic, focused responses
• —Higher values (1.0-2.0): More creative, varied responses

How It Works

1. 1.Planning Phase: AutoGPT uses Gemini to break down your task into actionable steps
2. 2.Execution Phase: Each step is executed sequentially with context awareness using Gemini
3. 3.Summary Phase: Results are compiled and presented in a readable format

Example Tasks

• —"Research and summarize the latest developments in AI"
• —"Create a Python script to analyze stock market data"
• —"Write a comprehensive blog post about sustainable energy"
• —"Design a marketing strategy for a new product launch"
• —"Analyze customer feedback and provide actionable insights"

Technical Details

• —Framework: Gradio 4.0+
• —API: Google Gemini API (google-generativeai)
• —Language: Python 3.11
• —Container: Docker-based deployment
• —Model: Google Gemini Pro (configurable)

Getting Started

1. 1.Clone this repository or use it as a template
2. 2.Set up your GOOGLE_API_KEY in Space Settings
3. 3.Deploy to Hugging Face Spaces
4. 4.Start executing autonomous tasks!

License

MIT License