All projects

Agentic chat

Agentic chat is an AI workspace that combines document search, memory, tools, and research flows so it can answer with better context and lower repeat cost.

Agentic chat was built as an AI workspace that can choose the right path for each request instead of treating every prompt the same way. It supports web search, reading URLs, working with documents, Google Workspace actions, and deeper research flows.

How It Was Built

The main technical choices behind the product, from system design to the parts that make it work day to day.

  • Built a routing layer that decides whether a request should use memory, documents, images, tools, or a mix of them.
  • Added document search with ingestion, chunking, embeddings, semantic search, reranking, and caching so similar requests do not always start from zero.

Why It Mattered

  • Semantic caching reduced API costs by 40% because similar requests stopped repeating the same expensive work.
  • Answers improved because the product chooses the right context for each request instead of pushing everything through one path.

What Stands Out

  • Treated agent behavior as a system design problem, not a prompt trick.
  • Used routing and caching to improve both quality and cost.

Tech stack

Next.jsTypeScriptOpenAIMem0PostgreSQLLangGraphRAGAgentic ResearchAgents

More projects

Additional work across AI products, developer tooling, and full-stack systems.

Browse all