— Open source · Local-first · AI-powered

Your knowledge,
under one dome.

Collect your notes, PDFs, videos, and links in one place — then let Many, your AI agent, search, explain, and connect everything. Local or cloud, your call.

  • Free & open source
  • macOS · Windows · Linux
  • Works offline
Dome · Many
Many
Hi, I'm Many. Ask me about your library →
Try asking —

✦ You've met Many — now meet your library.

100% Open source
10 AI providers
3 Platforms
Local-first Your data stays yours

— Many · Dome-native agent

A reasoning agent, not a chat window — with memory, tools, and your library on hand.

Many runs on the Dome-native agent runtime (the @dome/agent-core harness): persistent JSONL sessions, a context budget with live token estimates, multimodal attachments, web search, MCP tools, and human-in-the-loop resume. Use OpenAI, Anthropic, Google, Ollama, or OpenRouter — run local or in the cloud, Many cites where it can.

Many · context 128K of 200K · 64%
  • System 18K
  • History 54K
  • Tools 22K
  • Retrieved 28K
  • Attachments 6K
ContextUsageIndicator — donut + segmented budget. Push notifications on every token stream. Click the donut to cycle.
Many

"Your library is indexed. What should we find first?"

Dome · Many
you
Summarise chapter 4 of the dissertation corpus, and cite every claim with the page number.
many · reasoning
I read 1,204 pages across 6 sources. Chapter 4 makes three core claims, each grounded:
  1. Habits are not behaviors but context-bound attentional frames p.87.
  2. Long projects fail at the *handoff* between sessions, not at the work itself p.91.
  3. The 3-day rule is a measurable re-entry signal p.94.
4 tool calls · 2.3s
  • resource_search · "habits chapter 4" → 6 hits
  • retrieve_pages · 87–94 → chunks: 12
  • graph_neighbors · Smith 2024, Bohn 2017
  • draft_section · 412 words
you
Now turn that into 10 FSRS cards. Mix easy / hard / good.
many · thinking
calling generate_deck · waiting for FSM-2 schedule
Ask Many
128K / 200K · 64%
Try /summarise @chapter-4 #arxiv
/ @ # +
Fig. 02 — Many chat: reasoning agent with visible tool calls, FSRS-grade output, and a live context budget. Click Send to add a turn · click the donut to cycle budgets.

— Why Dome

See it work, not just read about it.

One calm workspace where your sources, your AI agent, and your study tools live together.

Everything in one place

Notes, PDFs, videos, web pages, and audio — organised into projects you can actually find again.

Ask, don't dig

Many searches your whole library and answers in plain language, citing your own sources.

Learn it, don't just store it

Turn any source into flashcards, quizzes, and study guides with spaced repetition.

Private by default

Run fully offline with Ollama, or bring your own cloud keys. Your library never leaves your machine unless you choose.

Dome · Library

All your sources · one place

  • 📝 Lecture notes — Week 1–6 Biology
  • 📄 Smith et al. (2024).pdf Thesis
  • Krebs cycle — video Biology
  • 🔗 arxiv.org/abs/2406… Reading
Summarise chapter 4 and cite it.
Three claims, each grounded — habits are context-bound p.87, projects fail at handoffs p.91.

Studio · generated from your sources

🧠 Mindmap ❓ Quiz · 12 🃏 Flashcards · FSRS 📑 Study guide 🗓 Timeline

Settings · privacy

  • Run locally with Ollama on
  • Cloud providers (optional) off
  • Library: ~/Dome · stays on this device 🔒

— How it works

See Dome work, step by step.

Dome · Library

Your library · 4 sources

  • 📝 Lecture notes — Week 1–6 indexed
  • 📄 Smith et al. (2024).pdf indexed
  • Krebs cycle — YouTube transcribed
  • 🔗 arxiv.org/abs/2406… adding…
What connects my Biology and Chemistry notes?
Both cover enzyme kinetics. Smith (2024) links the Krebs-cycle video to your notes on activation energy p.91.

Card 7 / 24 · due now

What is FSRS?

Click to reveal

A free spaced-repetition scheduler that times each card to how well you know it.

  1. 01

    Add your sources

    Drop in notes, PDFs, videos, and links.

  2. 02

    Ask Many

    Search and chat with your library in plain language.

  3. 03

    Learn & export

    Flashcards, study guides, and cited answers.

— Loved by people who live in sources

Built for researchers, students, and writers.

Dome · Community
PR
Dome replaced three apps in my research workflow. Everything's finally in one place.
PhD researcher
CS
I ask Many to summarise a paper and it cites the exact page. A lifesaver before exams.
Computer Science student
IW
Local-first means my notes stay mine. That's why I switched.
Independent writer

Illustrative — your story could be here.

— FAQ

Ask Dome anything.

Tap a question — Many answers in real time, just like in the app.

Dome · Many
Many
Hi! I'm Many. Tap a question and I'll answer it.
Frequently asked —
Is Dome free?
Yes — Dome is open source and free for personal and educational use. Commercial use needs written permission (see the license in the repo).
Does my data stay private?
Yes. Dome is local-first: your library lives on your machine. Run fully offline with Ollama, or add your own cloud API keys — your choice.
Which AI providers can I use?
OpenAI, Anthropic, Google, Ollama, OpenRouter, DeepSeek, Moonshot, Qwen, GitHub Copilot, and MiniMax.
What can I import?
Notes, PDFs, web pages, YouTube videos, audio, and images — all organised into projects.
Which platforms are supported?
macOS, Windows, and Linux.

Put it all under the dome.

Free, open source, and yours. Download Dome for macOS, Windows, or Linux.