Build agents that remember.
Opinionated patterns for shipping memory-aware AI. Each guide covers what to store, how the agent uses it, and the exact code to paste into your stack.
Building your first memory-aware chatbot
4 steps. About 10 minutes.
Model first, memory second. Prove out tone with a static prompt, then wire in Threadline at the system layer.
Install the SDK
npm install threadline-sdkInject context
tl.inject(userId, prompt)Update on each turn
tl.update({ ... })Expose trust dashboard
/accountConversational
Agents that talk to users in real time.Productivity
Tools that build on what they learned last session.Vertical
Domain-specific assistants with deep context.More ideas
Patterns that work the same way — guides coming soon.Vibecoding Platform
Remembers your stack, project structure, and preferences across every session.
LLM Wrapper / AI Product
Slot Threadline in at the system prompt level. Personalisation without infrastructure.
Multi-Agent Pipeline
Context travels with the user across every agent in your pipeline.
AI Research Assistant
Remembers topics explored, sources used, and open questions across sessions.
AI Legal Assistant
Knows the client's case history, preferences, and prior advice.
AI Mental Health Companion
Remembers the user's journey, mood patterns, and coping strategies over time.
AI Travel Planner
Knows your travel style, past trips, budget, and preferences.
AI Personal Assistant
Builds a long-term picture of routines, contacts, and recurring tasks.