Welcome to MLStack Documentation

Deploy LoRA adapters and AI agents with zero infrastructure.

MLStack is a platform that lets you upload fine-tuned LoRA adapters and run inference instantly. No servers, no containers, no infrastructure to manage.

Features

✨ LoRA Adapter Deployment
Upload your fine-tuned adapters and start running inference in seconds.

🤖 AI Agents with ReAct Loop
Build intelligent agents that can use tools, make decisions, and complete complex tasks.

📋 Structured Output
Define schemas for guaranteed JSON output from your models.

⚡ Serverless Inference
Pay only for what you use. No idle compute costs.

🔐 API-First Design
REST API with rate limiting, authentication, and comprehensive error handling.

Supported Base Models

MLStack currently supports the following models:

• meta-llama/Llama-3.2-1B-Instruct
• meta-llama/Llama-3.2-3B-Instruct
• meta-llama/Llama-3.1-8B-Instruct
• mistralai/Mistral-7B-Instruct-v0.3
• Qwen/Qwen2.5-14B-Instruct

Quick Links

• 🚀 Getting Started - Set up and make your first request in 5 minutes
• ⚡ LoRA Adapters - Train, upload, and deploy fine-tuned models
• 🤖 AI Agents - Build autonomous agents with tool access
• 📋 Output Schemas - Guarantee structured JSON output
• 📚 API Reference - Complete endpoint documentation

Why MLStack?

No Infrastructure Management

Forget about setting up servers, managing GPU instances, or scaling infrastructure. Just upload your adapter and start running inference.

Cost-Effective

Pay per inference with transparent pricing. No minimum commitments or idle compute costs.

Developer-Friendly

Simple REST API, comprehensive documentation, and code examples in multiple languages.

Next Steps

1. Create an account and get your API key
2. Train and upload your adapter or use a public one
3. Run your first inference via API or playground
4. Build an agent to automate complex workflows

Need Help?

• Check our API Reference
• Contact support at [email protected]