{
  "id": 11304138,
  "title": "Sity — AI assistant with persistent memory and its own cognitive architecture.",
  "url": "https://urgent.news/2026/10/01/sity-ai-assistant-with-persistent-memory-and-its-own-cognitive",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-01T23:09:26.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/sralex16/sity-ai-assistant-with-persistent-memory-and-its-own-cognitive-architecture-26fc"
  },
  "original_language": "en",
  "account": "Over the course of approximately four months, a self-hosted conversational AI named Sity has been developed. This assistant is capable of running entirely on a Raspberry Pi 4B, managing its backend, database, text-to-speech, and speech-to-text functions locally. The core of Sity's functionality is not simply a prompt with personality instructions, but rather a comprehensive cognitive pipeline. This pipeline processes each conversation turn through several stages: perception, emotional appraisal, goal tracking, episodic and semantic memory retention, and decision-making based on utility. The AI maintains episodic and semantic memory that persists across sessions, allowing it to remember details and track expectations over time.\n\nIt also features a utility-based decision engine capable of selecting from 10 different actions. Moreover, Sity's conversational AI has a reflective component and a self-model that evolves over time, enabling it to learn and adapt to user interactions. The relationship between Sity and its user is dynamic, as the AI can take initiative and follow up on matters it deems important.\n\nThe self-hosted platform for Sity involves running on a Raspberry Pi 4B equipped with 4GB of RAM, utilizing an SQLite database without any external dependencies. Text-to-speech is performed using Piper TTS, while speech-to-text is handled by faster-whisper STT, both of which operate locally. A Caddy reverse proxy and Cloudflare Tunnel facilitate remote access to the assistant. Sity is currently available in a public alpha stage, open for testing at https://sity.aletm.com/ and available on GitHub at https://github.com/SrAlex16/sity. The creator, SrAlex16, is particularly interested in receiving feedback from individuals who have built similar conversational AI systems, eager to compare architectural approaches.",
  "summary": "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…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}