Settings & AI Configuration

Inspect model status, configure open-weight endpoints, and manage local browser data.

Inference Engine Status

Current operational mode of LazyFlow AI

Current Mode:

Demo Mode

Target Model:

Local Heuristics Engine

Endpoint Host:

Local (Offline safe)

Configuring Real AI Inference

Connect an open-weight model via any OpenAI-compatible provider

LazyFlow AI connects to any OpenAI-compatible chat completions endpoint (such as Together AI, Groq, Ollama, vLLM, or OpenRouter) supporting open-weight models like **Qwen 2.5** or **Gemma 2**.

# Set these in your .env.local file or Vercel Environment Variables

AI_API_BASE_URL=https://api.together.xyz/v1

AI_API_KEY=your_private_api_key

AI_MODEL=Qwen/Qwen2.5-72B-Instruct

API keys are strictly managed on the server. They are never transmitted to client browsers or exposed in client bundles.

Privacy & Local Storage

How LazyFlow AI handles your documents, notes, and activity

• Browser-Local Persistence: Your automation jobs and activity history are stored in your browser's localStorage. This data belongs exclusively to your device and is never automatically synced to external databases.

• No Unsolicited Tracking: We do not log or retain document text on our servers. When an external AI provider is configured, request payloads are sent directly to that endpoint and are not saved in server files.

• Storage Boundaries: LocalStorage is not encrypted at rest by the browser. Avoid storing unencrypted high-security credentials in notes.