Lumino: Browse and Build with AI Agent

A free AI assistant for deep research, knowledge management, web replication, browser automation, etc. — powered by your own LLM.

🔥 Lumino is your capable AI assistant — powered by your own LLM. Open the side panel, and it actually acts. Lumino doesn't just chat about a page — it reads, clicks, fills, scrapes, saves files, and talks to your Obsidian vault, chaining built-in tools in a self-driven loop until the work is done. Bring any OpenAI-compatible LLM — point Lumino at OpenAI, Claude, Gemini, a local model, or any compatible provider. You stay in control of the model and your data. 💪 Key Features: 👉 Work with Purpose-Built Agents ✔️ Deep Research: Browses pages, extracts content, follows links, and synthesizes findings into a clear report. ✔️ Knowledge Management: Read, write, and organize notes in your local Obsidian vault via the Local REST API — your second brain, always within reach. ✔️ Web Replication: Scrapes a page's full DOM structure, computed styles, and assets so you can study or recreate any design. ✔️ Browser Automation: Fills forms, clicks elements, presses keys, and inspects the DOM to drive multi-step web flows on your behalf. ✔️ Custom Agents: Create your own agents with a tailored system prompt and a curated tool subset. Custom agents sync across your devices. 👉 Real Tools ✔️ Browser interaction — navigate, click, fill forms, press keys, inspect elements, take screenshots, scroll, close tabs ✔️ Page reading — extract visible text or structured content from any tab ✔️ Page scraping — capture DOM structure, computed styles, and resources; download files ✔️ File system — a full OPFS workspace with ls, read, write, edit, glob, grep, rm, and export to your computer ✔️ Obsidian integration — list, read, and write notes in your local vault 👉 Chat That Gets Things Done ✔️ Multi-session: Create, switch, search, and delete conversations — each tied to the agent you used. ✔️ Step-by-step transparency: Watch every tool call and result as it happens. ✔️ Tab locking: Once the agent picks a tab, your tab-switching won't disrupt its work. ✔️ Markdown rendering with sanitization for safe display. 🔒 Privacy First ✔️ No telemetry, no analytics, no tracking — Lumino never collects or transmits your data to the developer. ✔️ API keys stay local — stored in chrome.storage.local, never synced to your Google account. ✔️ You choose the LLM — Lumino talks only to the OpenAI-compatible endpoint you configure. ✔️ Chat history and files live on your device, in IndexedDB and the Origin Private File System. 🎨 More ✔️ Three themes — Paper, ORYZO, and Fictional ✔️ Bilingual — English and Chinese with runtime switching ✔️ Free to use — bring your own LLM key, no subscription required 💻 How to Use: 🔸 Click "Add to Chrome" and pin Lumino to your toolbar. 🔸 Open the Options page and configure your LLM provider (base URL, API key, model name). 🔸 Optionally enable Obsidian integration via the Local REST API plugin. 🔸 Open the side panel, pick an agent mode, and start a conversation. ❓ FAQ: 📌 Do I need an OpenAI account? - No. Lumino works with any OpenAI-compatible LLM endpoint — OpenAI, Anthropic, Google, a local model, or any compatible provider. You bring your own key. 📌 Is Lumino free? - Yes. Lumino itself is free. You only pay for the LLM API you choose to use, if any. 📌 Does it work in my country? - Yes. Since Lumino connects to the LLM provider you configure, availability depends on your chosen provider, not on Lumino. 📌 Is my data safe? - Yes. Lumino stores everything locally — API keys, chat history, and files never leave your device except for the requests you explicitly send to your own LLM provider. 📪 Contact: Have questions or suggestions? Open an issue on our GitHub repository. See the full privacy policy at: https://raw.githubusercontent.com/the-invulnus/lumino/main/PRIVACY.md