Ancestral Brain runs entirely on your machine — no cloud, no subscriptions, no surveillance. It indexes your life, learns who you are, and connects to any AI tool with one copy-paste.
Free 14-day trial · Then $79 one-time · macOS & Windows
Scans your hardware in seconds. Recommends the best local AI model you can actually run. Downloads and launches it — ready to chat in under 10 minutes.
Indexes your notes, documents, and chat exports — with your consent, folder by folder. Builds a private knowledge base that answers from your life, with citations.
Runs a local MCP server. Copy one snippet into Claude, Cursor, or OpenClaw — now any AI can search your Brain. One paste. Every tool, instantly smarter about you.
An optional offline library: Wikipedia, WikiHow, iFixit, Gutenberg — ready when the internet isn't. Download ZIM packs you care about and search them locally.
Ancestral Brain is designed to be the easiest path to a private, personal AI. No terminal commands. No config files. No cloud account.
Download the app. It scans your CPU, RAM, and GPU in seconds — then recommends the best model your hardware can run well.
One click. The model downloads directly to your machine — no account, no API key. Bundled with a high-performance inference engine that uses Metal (macOS) or CUDA (Windows).
Tell Ancestral Brain which folders to index. It shows you exactly what it found and asks permission before touching anything. No secrets, no hidden scanning.
Go to Connect and copy your personal snippet. Paste it into Claude Desktop, Cursor, or OpenClaw's config. Now every AI you use can search your Brain — privately, locally.
Export a Brain Pack — a single zip containing everything Ancestral Brain knows about you. Move it to a new machine, share it selectively, or import it anywhere. You own it.
Ancestral Brain recommends the right model for your specific hardware — automatically. But here's what to expect:
| Machine | RAM | GPU / Chip | Experience | Recommended model size |
|---|---|---|---|---|
| Mac (M1/M2/M3/M4) | 16 GB+ | Apple Silicon (shared) | Excellent | 7B–13B models, fast |
| Mac (M1/M2 base) | 8 GB | Apple Silicon (shared) | Good | 3B–7B models, usable |
| Windows PC w/ NVIDIA GPU | 16 GB+ | RTX 3060+ (8 GB VRAM) | Excellent | 7B–13B fully on GPU |
| Windows PC w/ NVIDIA GPU | 16 GB+ | GTX 1080 / RTX 2060 (4–6 GB VRAM) | Good | 7B split CPU+GPU |
| Windows / Mac (CPU only) | 16 GB+ | None / integrated | Usable | 1B–3B, slower but works |
| Any machine | 8 GB | CPU only | Minimal | 1B–3B only |
Minimum to run: 8 GB RAM, 15 GB free disk space, macOS 13+ or Windows 10 64-bit.
Ancestral Brain scans your hardware at startup and recommends the best model automatically — you don't need to pick from this table manually.
No subscription. No per-user fees. No data sent anywhere. Pay once, own it.
Early waitlist members get launch-day pricing. We'll email you when it's ready.
Be first in line. Early members get launch-day pricing and priority support.
We store your email. Nothing else. No spam. Unsubscribe anytime.