Hindsight Cookbook

A collection of example applications and notebooks demonstrating how to integrate and use Hindsight.

Applications

Full-featured example applications demonstrating Hindsight integration patterns:

• chat-memory - Conversational AI with per-user memory
• crewai-memory - CrewAI agents with persistent long-term memory
• pydantic-ai-memory - Pydantic AI agent with persistent long-term memory
• deliveryman-demo - Interactive delivery agent with memory-based navigation
• hindsight-litellm-demo - Side-by-side comparison of memory approaches
• hindsight-tool-learning-demo - Learning tool selection through memory
• hermes-memory - Hermes agent with persistent long-term memory via Hindsight
• openai-fitness-coach - Fitness coach with OpenAI Agents and Hindsight memory
• sanity-blog-memory - Syncing Sanity CMS content to Hindsight
• chat-sdk-multi-platform - Cross-platform Slack + Discord bot with shared memory (Vercel Chat SDK)
• stancetracker - Track political stances using AI-powered memory

Notebooks

Interactive Jupyter notebooks demonstrating Hindsight features:

Core Tutorials

• 01-quickstart.ipynb - Basic operations: retain, recall, and reflect
• 02-per-user-memory.ipynb - Pattern for per-user memory banks
• 03-support-agent-shared-knowledge.ipynb - Multi-bank architecture for support agents
• 04-litellm-memory-demo.ipynb - Automatic memory with LiteLLM callbacks
• 05-tool-learning-demo.ipynb - Learning tool selection through memory
• 08-llamaindex-react-agent.ipynb - LlamaIndex ReAct agent with long-term memory

Quick Demos

• fitness_tracker.ipynb - Fitness coach with workout and diet memory
• healthcare_assistant.ipynb - Health chatbot demo
• movie_recommendation.ipynb - Personalized movie recommender
• personal_assistant.ipynb - General-purpose assistant with long-term memory
• personalized_search.ipynb - Context-aware search agent
• study_buddy.ipynb - Study assistant with spaced repetition

Getting Started

Prerequisites

Start Hindsight using Docker:

export OPENAI_API_KEY=your-key

docker run --rm -it --pull always -p 8888:8888 -p 9999:9999 \
  -e HINDSIGHT_API_LLM_API_KEY=$OPENAI_API_KEY \
  -e HINDSIGHT_API_LLM_MODEL=o3-mini \
  -v $HOME/.hindsight-docker:/home/hindsight/.pg0 \
  ghcr.io/vectorize-io/hindsight:latest

• API: http://localhost:8888
• UI: http://localhost:9999

See Hindsight documentation for other LLM providers.

Running Notebooks

Each notebook can be run independently. Install dependencies:

cd notebooks
pip install -r requirements.txt
jupyter notebook

Running Applications

Each application has its own setup instructions in its README.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.

License

MIT