Abah_chat – Persistent Memory Local Chatbot John Abah released Abah_chat, a local-first conversational AI tool with persistent memory that stores chat histories across sessions and allows users to switch between local models without losing context. The tool, powered by Ollama, requires Python 3.12.13 and is available on GitHub, enabling users to resume conversations and query past discussions. A local-first conversational AI tool featuring persistent memory. Return to your chat history at any time or seamlessly switch models while retaining full context of your previous conversations. Overview -overview Key Features -key-features Prerequisites -prerequisites Installation -installation Usage -usage Switching Models -switching-models Credits -credits Contacts -contacts Buy Me a Coffee -buy-me-a-coffee abah chat is designed to eliminate session loss during local LLM interactions. It stores conversation histories across sessions in persistent storage, allowing you to: - Resume conversations exactly where you left off. - Switch between different local models without losing chat context. - Query the LLM about past discussions and saved history. Persistent Chat History: Conversations are saved locally across restarts. Model Interoperability: Swap the underlying LLM at any point and continue referencing prior context. Context Querying: Prompt the LLM directly to summarize or recall where you left off in past sessions. 100% Local & Private: Powered locally through Ollama. Python: 3.12.13 or Python 3.12+ Ollama: Installed and running locally - Clone the repository and navigate to the project root: git clone https://github.com/john-abah/ABAH CHAT.git cd ABAH CHAT - Install the package in editable mode: pip install -e . - Pull the model in Ollama: ollama pull gemma4:12b - Start the Ollama background service: ollama serve - Launch the chat assistant in your CLI: abah chat - Interact with the assistant: - Type your messages normally to chat. - Ask contextual questions such as "Where did we leave off last time?" or "Summarize our previous conversation" . If you wish to use a different model pulled from Ollama: - Pull your desired model: ollama pull