Local browser LLM
Run supported browser-native models without a separate local service.
Follow the path in order: install the extension, choose how AI should run, then connect the place where your knowledge will live.
Use the Chrome Web Store when available. If it is blocked or unavailable, download the ZIP release, unzip it, then load the folder through Chrome extensions developer mode.
Pick the AI path before routing captured content into a destination. This keeps privacy expectations clear from the start.
Run supported browser-native models without a separate local service.
Connect OmniFlows to local models such as Llama, Qwen, or Mistral.
After AI is configured, choose where OmniFlows should persist highlights, summaries, tags, and reading trails.
Sync browsing trajectories and captured knowledge to your vault.
Structured workspaceStore captures in a Notion database with organized properties.
Keep data and configuration backed up in your personal cloud.
When setup fails, check the layer that changed most recently: extension install, AI provider, destination permissions, then local ports or API keys.