Just released pact-langchain v0.1.0 - a LangChain memory system that tracks emotional context alongside conversation history.
Background: 25 years in behavioral science. Built this because AI conversations often miss emotional nuance. When users say "I'm frustrated," most systems just store the text without understanding the emotional state that shapes the conversation.
PACT adds emotional intelligence to LangChain's memory system:
- Tracks emotional state across conversations
- Session-based persistence
- Cloud-native architecture
- Drop-in replacement for existing LangChain memory
This is an MVP (v0.1.0). Basic emotional tracking works, but advanced analysis is coming in v0.2.0.
Technical stack: Python, FastAPI, SQLAlchemy, deployed on Render.
Would love feedback from the HN community - especially on:
1. Emotional analysis approaches (LLM-based vs traditional ML)
2. Privacy considerations for emotional data
3. Use cases that need this
You're seeing something real in all three observations:
AI narrative consistency: Yes, this is infrastructure for maintaining coherent identity over time — whether human ↔ AI or AI-agent organizations.
Niche packaging: Agreed. While the core is general infrastructure, entry strategy should be vertical-specific (customer support, coaching, education, etc.)
Writers/world-builders: Great callout. Character consistency across long narratives is exactly the kind of "emotional continuity" problem StoryKeeper addresses.
The broader vision is relational infrastructure for AI systems. The go-to-market is specific solutions for specific communities. Writers might be a great early adopter group
Fair feedback. Claudesidian is a productivity system where you organize knowledge and Claude assists. StoryKeeper is relational infrastructure that maintains emotional continuity across AI sessions and agent handoffs. Different layers of the stack, both valuable. I'll update the docs to make this clearer — appreciate the push for concreteness.