Sity — AI assistant with persistent memory and its own cognitive architecture.

Alex — Thu, 01 Oct 2026 23:09:26 +0000

I've been building Sity for the past ~4 months. It's a
self-hosted conversational AI that runs entirely on a
Raspberry Pi 4B (backend, database, TTS/STT — all local).
The language model is Claude Haiku via API, but everything
else is on-device.

What makes it different from a typical AI assistant:

It has a proper cognitive pipeline — not just a prompt with
personality instructions. Each conversation turn goes through:
Perception → emotional appraisal → goal tracking
Episodic and semantic memory that persists across sessions
A utility-based decision engine (10 possible actions)
Reflection and self-model that update over time

The relationship with the user evolves. It remembers things,
tracks expectations, and can take initiative when it detects
something worth following up on.

Self-hosted details:
Runs on Pi 4B (4GB RAM)
SQLite database, no external dependencies
Piper TTS + faster-whisper STT (both local)
Caddy reverse proxy + Cloudflare Tunnel for remote access
AGPL-3.0

Current status: public alpha, open for testing

🌐 https://sity.aletm.com/
💻 https://github.com/SrAlex16/sity

I'm particularly interested in feedback from people who've
built similar things — curious how the architecture compares
to other approaches.

— SrAlex16

DEV Community: Alex

Sity — AI assistant with persistent memory and its own cognitive architecture.