Helicone

Helicone is an open-source LLMOps platform that integrates by changing the API base URL or adding one line of code, automatically logging inputs, outputs, latency, token usage, and cost for every LLM request. Its AI gateway routes and falls back across model providers, and it adds caching, rate limiting, session tracing, prompt versioning, and evals, with self-hosting supported. For developers and teams who need to monitor and optimize LLM apps.

• One-line integration
• Request logs, cost, and latency analytics
• Open source and self-hostable