HeliconeUnclaimed

Helicone is a platform for teams building applications on large language models. It sits between an app and its model provider, routing requests and recording what happens to each one, so developers can see how their AI features behave in production.

The page describes several parts. A gateway routes requests and offers caching, rate limits and automatic fallbacks. Monitoring covers requests, sessions, users and custom properties, with a query language called HQL, plus alerts and reports. A prompts area includes a playground, datasets and scores for testing changes. Webhooks and data export are also listed.

Setup is described as a one line integration. Helicone names OpenAI, Anthropic, Azure, LiteLLM, Anyscale, Together AI and OpenRouter among the providers it works with, and lists an Other option for the rest.

The product is aimed at businesses, and the page says it is used by fast-growing AI companies. The Team plan lists SOC-2 and HIPAA compliance, and the Enterprise plan lists SAML SSO and on-premise deployment.

The Hobby plan is free and includes 10,000 requests and 1 GB of storage for one seat. Pro costs $79 per month and includes unlimited seats, alerts and reports. Team costs $799 per month. Both paid plans come with a 7-day free trial, and usage-based pricing applies on top of the plan price. Enterprise pricing is by contact. Startups under two years old with up to $5M in funding can get 50% off the first year. Open-source companies can get a $100 credit for the first year, and students can use the platform for free.