Zine is built for individuals. If you manage customer relationships across a team — CS, sales, or account management — Dossium is the B2B version: customer context that compounds across your whole organization.
Learn about Dossium →You open Claude.ai. Or ChatGPT. Or Gemini. Or T3Chat — maybe your favorite because it lets you hop between models mid-conversation.
You ask: "What did Sarah say about the budget in that call last week?"
The AI says: "I don't have access to your meetings or emails."
So you switch tabs. Search your email. Pull up the transcript. Copy the relevant part. Paste it back. Try again.
This happens constantly. The AI tools are genuinely impressive. But they're impressive in a vacuum. They don't know who Sarah is. They don't know what call you mean. They don't know anything about your actual life.
We built Zine to fix that.
The Problem With Every AI Chat App
There are more AI chat apps than anyone can count now. Most of them are excellent at what they do: clean interfaces, fast responses, sensible defaults. A few — like T3Chat — have made the multi-model experience genuinely delightful. Switch between Claude, GPT-4o, and Gemini in the same thread. Pick the right model for the right question.
The model switching problem is solved. The interface problem is mostly solved.
The context problem is not solved.
Ask any of these apps about something real in your life — a specific conversation, a document you wrote last month, what someone said in a meeting — and they hit the same wall. They're brilliant generalists who know nothing about you specifically.
The workaround everyone uses is pasting. Paste the email, paste the transcript, paste the doc. This works, barely, for one-off tasks. It breaks down immediately for anything with history: a project that's been running for three months, a customer relationship with dozens of touchpoints, a research thread you've been building for a year.
You shouldn't have to paste your life into a chat window every time you want to think about it.
What We're Building
Zine is a personal AI chat app where every conversation starts with your actual context.
Connect your email, calendar, notes, files, and the tools you actually use. We keep everything synced automatically. Then when you chat — with Claude, GPT, Gemini, or any model you prefer — your questions get answered with your real information, not just the AI's general training.
The difference:
Without Zine: "What was that article I saved about sleep and memory?" → The AI says it doesn't have access to your saved links.
With Zine: "What was that article I saved about sleep and memory?" → The AI finds it in your Notion reading list, connects it to the notes you took after reading it, and surfaces the three other pieces you saved on the same topic. All sourced. All in context.
Same question. Completely different answer — because the AI finally knows what you know.
How It Works
Connect Once, Know Everything
Link your tools through OAuth: Gmail, Google Calendar, Slack, Notion, Google Drive, Dropbox, GitHub, and more. We sync everything automatically, in real time. New email? Indexed in seconds. New note in Notion? Processed and searchable immediately. New Slack thread? Part of your context before you've even closed the tab.
No manual uploads. No copy-pasting. No re-indexing. Connect once, and your context stays current.
Chat With Any Model
Pick the model that fits the task. Claude for nuanced reasoning. GPT for breadth. Gemini for speed. Grok, Deepseek, Mistral — they're all here. Switch mid-conversation without losing your thread. The model changes; your context doesn't.
Every model gets the same foundation: what you actually know. The result varies by model. The grounding doesn't.
Not Just Search — Understanding
Most "AI with your documents" tools are keyword search with a chatbot on top. You ask a question, they retrieve text chunks that contain matching words, the AI reads the chunks aloud. It's fine for simple lookups. It falls apart for anything relational.
Zine works differently. We extract entities — people, places, projects, topics — and build the relationships between them. Your friend Marcus isn't just a name that appears in your email. He's a person, linked to that group trip you planned in June, mentioned in a dozen Slack threads, connected to the reading list he sent you, and the person who introduced you to that podcast you've been binging.
When you ask about Marcus, you get Marcus — everything connected to him, across every source, in order.
Time Actually Matters
The rent your landlord quoted in October and the number on the new lease are different facts. The project plan from last week's notes is different from the one two weeks ago. What your professor said about the final before the syllabus change is different from the updated requirements.
Most systems don't track this. They retrieve a document, not a timeline.
Zine understands when things happened and how they changed. Ask "what's the latest?" and you get the latest — not the first mention from three months ago.
Your AI Remembers
Stop re-explaining your context every time you open a new chat. Zine builds a persistent memory from everything you connect. Start a conversation today about a side project, come back next week — it still knows the history. The people involved. The decisions you made. What changed since you last checked.
It's the difference between a brilliant assistant who just started and one who's been working alongside you for months.
Who This Is For
Zine is for anyone accumulating more than they can hold in their head.
If you're researching anything — a thesis, a market, a decision, an obsession — Zine lets you ask questions across hundreds of papers, notes, and tabs and get cited answers. Not search results. Actual answers, from your actual sources.
If you're in school, your lectures, readings, and notes stop being scattered files and become one searchable brain. Ask "what did the syllabus say about the final?" or "summarize everything I've read about this topic" and get a real answer.
If you're working on projects — side projects, creative work, anything with a history — Zine keeps the thread. Every file, every note, every conversation connected. Ask "where did I leave off?" and know immediately.
If you're a developer using Cursor, Claude Code, or Windsurf, Zine's MCP server gives your coding agent access to your actual project context: the architecture discussion from that Slack thread, the reasoning behind a PR, the spec doc in Notion. Your agent stops being generic and starts knowing your codebase.
The common thread: context you've built up over time, scattered across tools, finally available in one conversation.
Built on Three Years of Infrastructure
We didn't start with the chat interface. We started with the hard parts.
Three years ago, we began building Graphlit — the infrastructure platform that handles multimodal content ingestion, entity extraction, identity resolution, knowledge graph construction, and vector search. We built it for developers shipping AI-powered products. It's now powering applications across dozens of industries.
Zine is what happens when you build a consumer product on top of that infrastructure.
The connectors, the entity extraction, the temporal modeling, the hybrid search — all of it was already built. We could focus entirely on making the chat experience genuinely useful. That's a different problem than most AI chat apps are solving, and it's why Zine feels different.
We also built Dossium on the same foundation — a customer context platform for CS and account teams. Same infrastructure. Different interface, different scope. Zine is personal; Dossium is organizational. Both are grounded in the same belief: context that's structured, identity-resolved, and time-aware changes what AI can actually do.
Early Access
Zine is free to start. Connect your tools, and your first search is ready in minutes.
We're also opening up a Pro tier for people who want higher limits and priority access to new features. If you've been waiting for an AI assistant that actually knows your stuff, this is it.
If you want to talk through how Zine fits your workflow — or if you're building something on Graphlit and want to understand the underlying platform — I'm always up for a conversation.
For more on the underlying architecture: The Context Layer AI Agents Actually Need and Why Your Team's Knowledge Isn't Just Documents. For the customer context use case, see Introducing Dossium.