Blog
Notes from the team building xysq, the trust layer for AI agents and teams.
- Engineering
Case Study: The Price Changed in June. The Copilot Still Quotes March.
A sales copilot quotes a price that stopped being true two quarters ago. Here is why searching harder doesn't fix it, and what we do instead, with the output from a live run.
- Engineering
Case Study: A Customer Support Memory Layer in Two Vaults
Two memory stores, three lookups, masked at the door. The system wrote its own known-issues doc, and every code block ran against production.
- Engineering
Debugging Agents in Production: Context Engineering with HITL
You shipped an agent and it broke in production. The model was fine; the context was wrong. A field guide to debugging context engineering, prompt to queries to context to answer, and closing the loop so the same failure does not repeat.
- Product
When your star employee quits, the memory quits too
Your best engineer leaves and takes years of unwritten context with them. Here's how to set up a shared team memory so that context stays with the team, and get a new hire's AI up to speed on day one.
- Insight
Memory is the boring half of the self-improvement loop
Freeze a model's weights and it still gets better at a task? The learning went somewhere outside the weights. That somewhere is memory. A tour of Mem0, Letta, Zep, and Memora, and why a real self-improvement loop can't work without a store to compound into.
- Engineering
Give your team's AI a shared memory
Every dev on your team uses their own AI tool, and each one only knows what that one person told it. Here's how to set up a single shared memory so a fact saved once is known by everyone's AI.
- Insight
Skills Are Becoming the Real Infrastructure Behind Agentic AI
Most AI conversations focus on models and memory. But the quietest shift is happening at the skills layer — where agents learn not just what to do, but how.
- Engineering
Long-term memory for LangChain agents
Three LangChain agents — an intake nurse, a doctor, and a pharmacist — share one patient's memory through xysq. Same patient, three separate processes, one memory layer that survives restarts.






