We developed a a creative way to compress prompts without loosing the context. We're looking to stress test this and get some feedback. We believe that this wil

We developed a a creative way to compress prompts without loosing the context. We're looking to stress test this and get some feedback. We believe that this will reduce your token costs and latency as well. Currently, we're running it for free with 1K token/second but plan to release the 30K soon. It's easy to use, try the demo or the API. Please let us know your thoughts! Thanks =)