AI Chat with Memory

What It Does

Deploy a stateful AI chat endpoint that remembers every conversation. Unlike standard LLM APIs that forget context between requests, this proxy stores conversation history in a vector database and automatically retrieves relevant past interactions for every new message.

Your users get a ChatGPT-like experience — but powered by your knowledge base, your system prompt, and your rules.

How It Works

1. User sends a message to your proxy endpoint
2. Context retrieval — Aerostack searches past conversations and your knowledge base using vector similarity
3. LLM generation — the AI responds with full awareness of conversation history and your uploaded documents
4. Memory storage — the exchange is embedded and stored for future retrieval

Use Cases

• Customer-facing chatbots that remember returning users and their preferences
• Internal knowledge assistants where teams ask follow-up questions across multiple sessions
• Onboarding flows that pick up where the user left off, even days later
• AI copilots in SaaS products that build context about each user over time

Quick Start

Create a new proxy from this template, upload your knowledge base documents, and start sending messages. Conversation memory is handled automatically — no session management code needed on your end.