You know you read it. You can't find it.
The answer exists somewhere: an article, a PDF, an email thread. But it's buried in one of a dozen tools, and none of them can tell you which. You end up re-searching what you already knew.
Trace is an AI browser that quietly turns your tabs, documents, and conversations into one private intelligence layer, so every AI you use starts from what you already know.
Your knowledge is scattered across tabs, documents, and conversations. Every tool holds a fragment; nothing holds the whole picture.
The answer exists somewhere: an article, a PDF, an email thread. But it's buried in one of a dozen tools, and none of them can tell you which. You end up re-searching what you already knew.
Tabs, documents, notes, conversations: each captures a fragment of your thinking. Close the session and the connections between them are gone. Your context lives nowhere.
Saved links and chat histories pile up, but nothing connects them into your context. So every conversation with an AI begins with you re-explaining who you are, what you know, and what you're trying to do.
Full-text search across every page, PDF, and document you've ever opened. Describe what you remember, not the title, and Trace surfaces the exact one.
Tell the agent what you need and it takes over the page: filling forms, comparing prices, pulling data across tabs, while you keep working.
Ask for a document, a slide deck, a spreadsheet, even a landing page, and Trace builds it inline: a real, editable artifact, not just a wall of chat text.
Ask “which papers I've read compare RAG to fine-tuning” and Trace doesn't just search text, it walks the graph.
The smaller conveniences that add up: less typing, less choosing, less re-explaining.
Trace is building the personal and collective intelligence layer for the AI-native era. These are the four steps that get us there.
Every page, PDF, document, email, and conversation you open flows into one private, connected knowledge graph. Indexed automatically, stored on your device.
From that graph, Trace builds a deep understanding of your goals, expertise, and workflows. Ask anything: it answers from your context, not from scratch, and that same context is what starts letting the interface itself adapt to how you work.
Beyond answering questions: Trace assists proactively, automates tasks, and executes multi-step work with your full context in mind. And because it all runs on your device, you decide how far it goes.
A collaborative intelligence layer where people, teams, and AI agents securely share knowledge, so organizations preserve what they know instead of losing it when people move on, with every user in control of their own data.
Everything you need to know about Trace and how it protects your privacy.
Yes. Download, install, and use Trace forever at no cost. No telemetry, no subscription.
Yes. Trace builds a Personal Memory, a persistent set of natural language facts about you, extracted automatically from pages you browse, AI sidebar conversations, social profiles (LinkedIn, GitHub, Twitter/X, Bluesky, and more), and documents you upload. These facts are injected into every chat turn so the AI always knows your context, where you work, what you're researching, what matters to you, without you having to re-explain it each time. You can view, edit, or delete any fact from the Profile panel.
Trace works with Ollama, Azure AI Foundry, OpenAI, Anthropic, and Google Gemini. You can configure different providers for summarization, chat, and navigation planning independently in Settings, and switch any time.
In addition to web pages, Trace indexes PDFs, Word documents (.docx), PowerPoint presentations (.pptx), Excel spreadsheets (.xlsx), and EPUB e-books that you open in the browser. All are processed automatically, text extraction and AI summary, so they're fully searchable and accessible in the AI sidebar.
Yes. All your indexed pages, documents, and summaries are stored locally on your device. Nothing is sent to any external server unless you choose a cloud AI provider for inference.
Trace currently supports macOS (Apple Silicon and Intel) and Windows. Linux support is on the roadmap.
Trace indexes every page and document automatically, no manual saving. You search with natural language ("show me that PDF about Rust I read last month"), ask questions about what you read, and the AI builds a persistent memory of your interests from your conversations and browsing. Bookmarks require you to remember to save; Trace organizes automatically, so you don't have to.
It learns with you, works for you, and your data stays on your device, always.