Your entire knowledge base,
instantly askable.
A personal, grounded AI assistant for your second brain. Grounded in your own notes, with optional web search for the gaps.
Native integration with every frontier model, plus fully local
Bring your own API key — or run 100% offline with a local model.
Notes are where great ideas
go to hide.
You've spent years building a digital garden in Obsidian or Markdown. But when you actually need an answer, you're stuck digging through folders, chasing broken tags, and losing your flow. Your “Second Brain” has become a digital junk drawer.
You don't need more notes.
You need a way to talk to the ones you already have.
One keypress to remember everything.
QuickAI brings the power of a large language model to your local file system. Your notes aren't used to train anything, and we never store your content. Just instant recall of every thought you've ever recorded.
How it works
Five primitives,
one keypress.
Every answer traces back to a file.
Grounded, not guessed. Every claim ships with a clickable link to the exact markdown file it came from.
Today's chat
is tomorrow's context.
Save any chat back to your notes as a summarized markdown file. Your assistant's memory compounds.
Highlight, then
reply.
Highlight text anywhere on your Mac, hit Reply, drop it in as context.
We never store
your notes.
On-device by default. BYOK for frontier. We never store your content.
Your model,
your key.
Plug in GPT-5.4, Claude 4.7, Gemini 3.1 — whatever you already pay for. Or run fully offline with a local MoE model via Ollama. Your tokens stay in your keychain. We never see them.
- GPT-5.4OpenAIyour key
- Claude 4.7 OpusAnthropicyour key
- Gemini 3.1 ProGoogleyour key
- Llama 4 ScoutOllama · localoffline
- Qwen 3.6 AgentOllama · localoffline
Stop searching.
Start synthesizing.
This is the workflow you were promised. A silent partner that knows exactly what you know.
Roll out company-wide
in an afternoon.
One admin sets up your company knowledge, tools, and AI access. Everyone else just opens the app — no API keys, no setup, no IT ticket.
No API keys for staff
An admin connects one company provider key. Everyone else gets AI instantly through your managed gateway — nobody pastes a key or touches a setting.
Your company brain, shared
Company context, escalation rules, and tone are written once and inherited by every seat. New hires are productive on day one.
Plugs into your stack
Answer from your ERP (Frappe), tickets, Notion, Linear, and the web. Turn a ticket ID into an action — not a copy-paste.
Cost & control
See spend per person and per model. Choose which models your team may use. Manage seats as you grow.
From $120/seat/year · minimum 3 seats
Pricing
One-time payment. Yours forever.
Every tier is a lifetime license with one year of free updates. We never store your content.
Local-first Inference
Run fully offline via Ollama, or bring your own key when you want frontier power. We never route your notes through a server of ours — because we don't run one.
Instant Recall
Find answers across thousands of files in milliseconds. No cloud latency, no loading spinners.
Zero Subscriptions
Pay once, own it forever. We believe premium tools shouldn't be a rent-trap for your knowledge.
- Lifetime license
- 1 year of free updates
- Unlimited knowledge bases
- Unlimited Save to knowledge
- Multi-model picker
- Direct line to the founder
- Lifetime license
- 1 year of free updates
- Unlimited knowledge bases
- Unlimited Save to knowledge
- Multi-model picker
Unlocks when Early Bird sells out.
- Lifetime license
- 1 year of free updates
- Unlimited knowledge bases
- Unlimited Save to knowledge
- Multi-model picker
Unlocks when Early Bird sells out.
Common questions.
Everything you need to know about privacy, licensing, and local AI.
Is my data private?
QuickAI itself has no server — on a personal plan, nothing you do routes through us, and we never store your notes or conversations. Two modes: with your own API key, requests go straight to the provider you chose (OpenAI, Anthropic, Google) under your account — not us, and not used to train models. Or go fully on-device with a local model, where nothing leaves your Mac at all. Team plans add an admin-managed gateway (see the team questions below) — it proxies requests so staff don't juggle keys, and meters usage, but never stores your content.
Which AI models do you support?
Bring your own key for frontier models — GPT-5.4, Claude 4.7 Opus, or Gemini 3.1 Pro. That's what most people use, and your key stays in your Mac's keychain. If you'd rather stay fully offline, you can run Llama 4 Scout, Qwen 3.6, or Gemma locally via Ollama or LM Studio — that takes a bit more setup and a capable Mac, but nothing leaves the device.
What does 'Lifetime License' really mean?
You pay once, and the version of QuickAI you buy is yours forever. No monthly 'rent' for your own knowledge. This license includes one full year of feature updates. After that, you can keep using your version forever or renew for another year of updates.
Does it work with my existing notes?
Yes. QuickAI is a 'read-only' layer that sits on top of your existing folders. It works with Obsidian, Apple Notes, and any folder of Markdown or PDF files. No vendor lock-in, ever.
Do my teammates each need an API key?
No. On a Team plan, one admin connects a single company provider key; everyone else uses AI through your managed gateway and never touches a key. You choose which models are allowed, and see usage per person and per model.
Wait — doesn't QuickAI run locally?
For individuals, yes: fully local or your own key, and nothing leaves your Mac. Team plans add an optional admin-managed gateway so staff don't handle keys — your company key is encrypted and used only on your behalf, with per-seat usage you can see. You pick the model; individuals keep the local / BYO option.
How are team seats licensed?
Buy the seats you need (from $120/seat/year, minimum 3). The first person to activate becomes the admin and sets up the shared workspace; everyone else is provisioned automatically. Add seats as you grow.