How AI Works in Glassy
Understand how Glassy runs AI in your browser via WebGPU — two on-device runtimes, what they do, and when cloud AI is an option.
15 results
Understand how Glassy runs AI in your browser via WebGPU — two on-device runtimes, what they do, and when cloud AI is an option.
Deploy Glassy as a self-hosted single-user appliance with Docker Compose — four required environment variables and you're running in minutes.
Capture specific page regions or individual DOM elements with pixel-perfect precision — HiDPI awareness, offscreen rendering pipeline, and CSS selector persistence.
Discover how Glassy runs 384-dimensional embeddings in your browser via WebGPU to power Related Notes and Smart Tag Suggestions — no cloud, no data leaving your machine.
Understand the separation between your self-hosted appliance and the Glassy cloud — membership verification, local JWT auth, and the 30-day offline window.
Understand how the Glassy Companion offline queue works — persistent storage in chrome.storage.local, automatic 60-second flush, badge counts, and manual queue management.
Use Glassy's AI assistant to ask questions about everything you've saved — with answers grounded in your actual notes, bookmarks, and captures, complete with source citations.
Configure environment variables, change ports, set up CORS, and access your self-hosted Glassy from your phone or other devices via Tailscale, Cloudflare Tunnel, or Caddy HTTPS.
Route captures to different Glassy accounts from the same browser extension using the account picker — personal and work accounts side by side.
How Glassy Cloud routes AI queries to cloud providers when local models can't handle the job — and what that means for your privacy and costs.
Protect your self-hosted Glassy data with automated daily backups, optional AES-256-GCM encryption, manual backup and restore, and storage monitoring.
Bring your own API keys for OpenAI, Anthropic, and Gemini to bypass Glassy Cloud's AI cost caps and use the models you already pay for.
Keep your self-hosted Glassy updated with docker compose pull or auto-update with Watchtower, check container health, and diagnose common issues.
Run capable LLMs like Llama and Qwen on your self-hosted Glassy server via Ollama — no API costs, no data leaving your machine, optional GPU acceleration.
How AI billing works on the self-hosted Glassy appliance — no metering, no cost caps, no system cloud keys. BYOK and Ollama are your paths to AI.