SHIM

Hey Product Hunt! I’m Mertcan, one of the makers behind SHIM. Let’s be honest: Building with LLMs is terrifyingly easy but expensive and risky right now. We built SHIM because we faced two massive problems while developing AI for enterprises and our hobby projects: The "Bill Shock": We realized we were paying OpenAI for the exact same query thousands of times. A user in your startup tends to prompt similar queries to your AI. The "Compliance Nightmare": Lets be honest: We all pasted an users data to ChatGPT. We have realized this is a huge risk: One accidental paste of a customer's ID or credit card into ChatGPT, and we were facing a GDPR/compliance lawsuit. We needed a middleware that acts like an "Airlock" between our apps and the LLM providers. So, we built it. SHIM allows you to: Stop burning cash: Our Semantic Caching recognizes repetitive questions and answers them instantly from your local cache. Zero cost. Zero latency. In addition to this, our Smart Normalization truncates your prompts and saves you more tokens. Sleep better: Automatically detect and redact PII (Emails, Credit Cards, IDs) before they leave your server. Observability and no vendor lock-in: You can change your model provider on the fly and you can observe your usage easily via our dashboard. All of this with just one line of code change. If you are interested please reach out to me, we are looking for design partners to shape our product for the better.