Every feature, on every plan. You pay for the AI — never for unlocks. Bring your own AI and pay almost nothing, or let us supply everything.
The work you do live — chat, pipeline sessions, the coding workspace. You're in the room.
Recurring work you ordered — scheduled research, weekly reports, inbox deliveries. Your errands, on your schedule.
The brain that never stops — memory, connections between notes, your voice, the Morning Digest. Nobody orders it; it's the product working for you.
You connect the AI you already have — API keys, a Claude Code subscription, or models running on your own machine. Your conversations run on your AI, at your cost, with no markup from us, ever. The plan pays for the always-on brain we run on your behalf.
Cortex Buddy was built from day one to spend AI carefully — including yours. Every task is routed to the least expensive model that can do it well: quick classifications go to fast, cheap models; only genuinely hard work reaches the expensive ones. And you hold the dial — throttle between Cruise, Drive, Overdrive, and Full Throttle to decide how far your AI budget stretches versus how much firepower each task gets. Point the app at economical models and a few dollars covers a serious month of work. Your cost-effectiveness is a top priority whether we're supplying the AI or you are — we'd rather your dollars go far than burn fast.
The modes are explained under the AI Included plans below — same dial, wherever your AI comes from.
No API keys, no subscriptions elsewhere, nothing to set up. Everything — your conversations, your commissioned jobs, and the ambient brain — runs on us, against a weekly meter that resets every Monday. It feels unlimited until the meter fills. And you hold the throttle: run efficient models in Cruise and the meter stretches; run the strongest models in Full Throttle and it drains fast doing gorgeous work.
$0, no keys, no card. Everything runs on us using fast, efficient open models (US-hosted). Every feature works against a small weekly meter: enough to really use the app, not enough to live in the pipeline. Digest once a week, a couple of inbox subscriptions. When the meter fills, Monday isn't far.
You only ever see simple bars. Model and mode choices are your throttle — the same week of work can read 80% in Full Throttle or 12% in Cruise on efficient models. If the meter fills: wait for Monday, or opt in to overage (off by default, priced transparently, never a surprise).
Any one of these counts as "bringing your own AI" — connect it in Settings and the Bring Your Own AI plans apply. Mix and match as you like.
Bring a key from your AI provider of choice. Your conversations bill to your key at the provider's own prices — we add nothing.
Already pay for Claude Code? Install the Cortex Buddy Connector on your computer and your subscription powers your pipeline and coding workspace. Flat-rate AI you already own. Needs your machine on — an always-on Mac mini makes it 24/7.
Run models on your own machine — a Mac mini in the closet works. The Connector makes your hardware a provider. Your AI never leaves your house, and the compute is free. Offline hardware means that work waits for you, never silently moves to a paid path.
No buy buttons yet — we're onboarding beta users first. Leave your email and we'll send an invite when your spot opens. Beta users are grandfathered on launch pricing.
We email you once about the beta. No spam, unsubscribe anytime.
Draft pricing. All dollar figures are launch anchors and may be tuned from beta usage before general availability — beta users keep the pricing they joined on.