Helicone

Overview

Helicone is an AI Gateway and LLM observability platform for AI engineers. It routes, debugs, and analyzes AI applications, offering one-line integration to log requests from providers and frameworks. It supports monitoring, evaluation, experimentation, prompt management, and self-hosting.

Key Features

• AI Gateway: access 100+ AI models with one API key through the OpenAI API, with intelligent routing and automatic fallbacks.
• Quick integration: one line of code logs requests from OpenAI, Anthropic, LangChain, Gemini, Vercel AI SDK, and more.
• Observability: inspect and debug traces and sessions for agents, chatbots, document processing pipelines, and more.
• Analytics: track cost, latency, quality, and more, and export to PostHog in one line for custom dashboards.
• Prompt management: version prompts using production data and deploy prompts through the AI Gateway without code changes.
• Playground and fine-tuning: rapidly test and iterate on prompts, sessions, and traces, and fine-tune with OpenPipe or Autonomi.

Use Cases

• Debugging and monitoring agents, chatbots, and document processing pipelines.
• Tracking cost, latency, and quality across AI applications.
• Testing and iterating on prompts, sessions, and traces in a playground.
• Versioning and deploying prompts using production data.

Getting Started

• Get an API key by signing up and add credits.
• Update the baseURL in your code to the Helicone AI Gateway and add your API key.
• Use a supported model such as gpt-4o-mini, claude-sonnet-4, or gemini-2.0-flash, or any model from the Helicone models list.
• View your logs in the Helicone dashboard and access 100+ models through one API.

Deployment & Requirements

• Helicone can be used through its hosted service or self-hosted.
• For self-hosting, use the Docker Compose setup; clone the repository, enter the docker directory, copy .env.example to .env, and start services with the compose script.
• For Enterprise workloads, a production-ready Helm chart is available by contacting enterprise@helicone.ai.
• Manual deployment is not recommended; Docker or Helm should be used.
• Helicone comprises Web, Worker, Jawn, Supabase, ClickHouse, and Minio services.

Before You Adopt

• License: Apache-2.0. Review its terms before using, modifying, or distributing the project.
• Manual deployment is discouraged; Docker or Helm are the recommended self-hosting paths.
• The listed integrations may be outdated, so users should check current documentation for provider or framework support.
• The production-ready Helm chart for Enterprise workloads requires contacting Helicone for access.
• Helicone states SOC 2 and GDPR compliance for enterprise readiness.