Hey everyone! I'm the maker of SeeCost from Japan. I was running LLM calls across OpenAI, Anthropic, and Gemini in different apps, and every time I wanted to know my actual spend I had to open three different provider dashboards and mentally add things up. That got old fast. SeeCost fixes this with a small SDK that monkey-patches `fetch` on the server side. It silently reads usage metadata from LLM responses — model name, token count, estimated cost — and sends it to your dashboard in real time. Your app's response is never delayed or modified. Your API keys and prompt content are never touched. The SDK only sees what the provider already returns in the response body. To keep things transparent, we've made the tracker SDK fully open-source. You can audit exactly what data is being sent here: [https://github.com/ronpaou/seeco...] Setup takes about two minutes: 1. `npm install @seecost/tracker` 2. `npx @seecost/tracker init nextjs` (or express / hono / node) 3. Add your SeeCost API key to `.env` From that point, every LLM request shows up grouped by app and provider. Would love to hear what providers or alert types you'd want added next!

SeeCost

All your LLM costs:OpenAI, Anthropic, Gemini — on one screen

All your LLM costs:OpenAI, Anthropic, Gemini — on one screen