suryansh-singh-137/personna-ai — explained in plain English
Analysis updated 2026-05-18
Run a personal chatbot companion that remembers your name, hobbies, and stresses across sessions.
Store conversation facts as searchable embeddings so they can be recalled by meaning, not exact wording.
Keep an evolving profile summary of a user that updates automatically after each conversation.
Run an entire AI companion locally with no cloud database or Docker required.
| suryansh-singh-137/personna-ai | 0xallam/my-recipe | 0xhassaan/nn-from-scratch | |
|---|---|---|---|
| Stars | 0 | — | 0 |
| Language | Python | Python | Python |
| Last pushed | — | 2022-11-22 | — |
| Maintenance | — | Dormant | — |
| Setup difficulty | easy | moderate | moderate |
| Complexity | 2/5 | 2/5 | 4/5 |
| Audience | developer | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires a free Groq API key, runs fully locally with no Docker or cloud database needed.
Persona AI is a command line chatbot companion, nicknamed Suzy, built to solve a specific annoyance with most AI chat tools: they forget everything the moment you close the conversation. This project tries to give a chatbot real memory, so telling it your name, hobbies, or what is stressing you out today means it can bring that context back into a conversation next week, without you having to re-explain yourself every time. The memory works in four layers modeled loosely on how people remember things. A core layer holds permanent basics about the companion's identity and about you, like your name and age, which never change during a chat. A semantic layer stores individual facts pulled out of past conversations, such as a hobby you mentioned, saved in a way that can be found by meaning rather than exact wording, so asking about your hobbies later can surface something you never called a hobby at the time. An episodic layer stores a short summary of the overall feel of each full conversation, not just the facts inside it. Finally, a user profile layer rewrites everything known about you into a single paragraph after each session, so the assistant starts the next conversation already knowing the basics instead of waiting to look anything up. At the end of every session, the whole conversation is summarized and saved, new facts are folded into the running profile, and older stored memories are cleaned up to merge duplicates or resolve anything that contradicts newer information. Technically, the project is written in Python and uses LangGraph to manage the conversation flow, ChromaDB as a local database for storing memories as searchable embeddings, and Groq's hosted Llama 3.3 model to generate replies. Everything runs on your own machine with no cloud database or Docker required, and memories are saved to a local folder on disk. Getting started means cloning the repository, installing dependencies with the uv tool, adding a free Groq API key, optionally customizing the companion's name and personality in a JSON file, and running it from the command line. The README does not mention a license.
Persona AI is a locally run chatbot companion with a four-layer memory system, built with LangGraph and ChromaDB, that remembers facts, feelings, and past conversations across sessions.
Mainly Python. The stack also includes Python, LangGraph, ChromaDB.
Setup difficulty is rated easy, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
double-check against the repo, no cap.