Your AI. Your machine. Nothing leaves.
Engram is a local-first personal AI: it remembers what you tell it, answers questions grounded in those memories, and runs scheduled agents — entirely on your own hardware. Local Llama 3.1 via Ollama, encrypted vector memory in Qdrant, tamper-evident audit log. No cloud LLM, no telemetry, no call-home.
The demo above was recorded with Wi-Fi off the whole time. That's the point.
| 🚀 engram-os/engram-os | One-command install in the README |
| 🌐 engram-os.com | What it is, how it works, docs |
| 💬 Discussions | Questions, setups, ideas — say hi |
Every "local-first" claim in the README comes with the shell command that verifies it.
