I built a zero-backend, browser-only alternative to Open WebUI and heavy LLM stacks

Description and use

Lab is a lightweight, fully client-side LLM interface that removes the need for heavy backends and Docker setups, stores data in IndexedDB, and supports local Ollama models.

Problem

Open-source LLM chat interfaces typically require heavy backends, databases, and Docker setups, which is overkill for simple local or browser-based use cases.

Outcome

Lab eliminates the need for heavy backends, reduces resource usage, and provides fine-grained context control while remaining fully local and offline-capable.

Workflow

• 1. Install Lab via the live demo or clone the GitHub repo.
• 2. Configure local Ollama by setting OLLAMA_ORIGINS=* in the systemd service file and restarting.
• 3. Open Lab in a browser over HTTPS.
• 4. Connect to the local Ollama instance at http://localhost:11434.
• 5. Use Lab’s UI to chat, inspect token usage, pin messages, and switch providers.
• 6. Persist data locally in IndexedDB via Dexie.js.
• 7. Optionally install Lab as a PWA for offline use.

Evidence

Everything executes entirely on the client side in your browser.

Source

Local Storage via IndexedDB: Chats, model configs, system prompts, and API keys are persisted locally in your browser using Dexie.js.

Source