Hey, I’m True, Co-Founder of @Edgee
For the past few months, we’ve been working on a problem we’ve been seeing in production AI systems:
LLM costs do not increase linearly with usage, they increase with context.
As teams add RAGs, tool calls, long chat histories, memory, and guardrails, prompts get larger and token spending increasingly becomes the main hurdle.
So we created a token compression layer designed to run before inference.
<a href