# Why I built StreamCtx: The hidden context problem in every LLM app

> Source: <https://dev.to/streamctx/why-i-built-streamctx-the-hidden-context-problem-in-every-llm-app-2e9p>
> Published: 2026-06-13 11:50:05+00:00

Every LLM app I've built has the same broken pattern.

Request comes in - reconstruct context from scratch - call LLM - throw context away.

It's wasteful, slow and breaks at scale.

Most developers building ai app end up stitching together Redis, vector database and custom middleware just to give their app basic memory.

It's fragile. It doesn't scale. And every team reinvents the same glue code.

StreamCtx is a streaming context database built specifically for LLM applications.

Open Source. MIT/Apache 2.0 Licensed.

solo founder.

Feedback and beta users welcome!