Helicone
AI gateway and LLM observability platform with Apache 2.0 self-hosting, Ollama integration, request tracing, and cost tracking.
Helicone combines an AI gateway with LLM observability for engineers building agents, chatbots, and document processing apps. You can self-host the open source observability platform under Apache 2.0 or use the hosted service. It also integrates with Ollama for apps that run models locally. Its hosted gateway routes requests to external AI providers, so those requests leave your machine.
The gateway gives applications an OpenAI-compatible API for accessing models across providers, with routing and automatic fallbacks when a request needs another provider. Integrations include OpenAI, Anthropic, Azure OpenAI, AWS Bedrock, and Gemini. Helicone records requests and lets you inspect agent traces and sessions, helping you investigate failures across a sequence of calls rather than looking at each response alone.
Cost, latency, and quality tracking help teams compare application behavior and spending. A playground lets you test prompts against sessions and traces, while prompt management supports versioning with production data and deployment through the gateway without code changes. You retain access to your prompts.
For teams with an existing AI stack, Helicone connects with LangChain, LlamaIndex, LangGraph, and the Vercel AI SDK. It can export analytics to PostHog for custom dashboards and supports evaluation through RAGAS. Fine-tuning integrations connect to OpenPipe and Autonomi.
