Compresr.

We invested in Compresr. The company is building context compression for AI systems.

The work starts from a specific customer problem and carries new technical capability into a product people can use.

What they are building

Compresr is building question-aware compression that removes tokens an AI model does not need while preserving the evidence that determines the answer.

The result is lower inference cost and latency without forcing teams to rebuild their model or application stack.

Why we backed the founders

The strongest frontier companies make difficult technology useful in the real world. That requires product judgment alongside technical ambition.

We backed the Compresr founders because they are building through that full system rather than stopping at the demo.