Multi-agent platform for prompt optimization, decentralized storage, and synthetic data generation.
MachinaMachina gives you three AI agents that work together:
Prompt Agent - Optimizes your AI prompts to reduce tokens and improve outputs. Detects bias and inefficiencies through guided questioning.
Filecoin Agent - Handles decentralized storage on Filecoin network. Opens browser sessions, requests testnet tokens, and provides real-time education.
Synthik Agent - Powered by Synthik Labs API, generates high-quality synthetic datasets for any use case. Creates training data, test data, and human-readable content using advanced AI models.
git clone https://github.com/fozagtx/machina-labs
cd machina-labs
pnpm install
pnpm devSet up your .env file:
GOOGLE_GENERATIVE_AI_API_KEY=your_key
BRAVE_SEARCH_API_KEY=your_key
Web Search - Brave Search API integration for real-time information
Browser Automation - Puppeteer-powered web navigation for Filecoin testnet operations
Bias Detection - Identifies problematic patterns in prompts
Decentralized Storage - Permanent storage on Filecoin network
Synthik Dataset Generation - Custom synthetic data for ML training and testing powered by Synthik Labs
The Prompt Agent uses Socratic questioning to help you identify inefficiencies in your prompts. Instead of giving direct answers, it guides you to discover optimization opportunities yourself.
The Filecoin Agent opens browsers to help you navigate Filecoin hosting, automatically requests testnet tokens (USDC & tFIL), and teaches you about decentralized storage.
The Synthik Agent leverages Synthik Labs' advanced API to create realistic datasets based on your specifications, adapting to any domain or industry requirements with high-quality synthetic data generation.
Built on Mastra AI framework with Google Generative AI, Filecoin network integration, Brave Search API, and Puppeteer for browser automation.
- Reduce AI API costs through prompt optimization
- Store optimized prompts permanently on decentralized storage
- Generate training data for machine learning models using Synthik Labs
- Create realistic test datasets for applications with Synthik's synthetic data engine
- Learn prompt engineering through guided practice
The platform includes automated documentation navigation - agents can open relevant docs based on your requests.
- Synthik Labs API Documentation - Complete API reference for synthetic data generation
