# Helicone An open-source AI gateway and LLM observability platform for developers building AI apps. - Findings24: - Site: - Preços: Freemium - Categorias: Gateways e roteamento de IA, Observabilidade de IA ## O que é Helicone? Helicone is an open-source AI gateway and observability platform for applications that use large language models. It helps developers route model requests, investigate application behavior, and understand usage, latency, and costs across multiple AI models. Teams can integrate Helicone into existing AI workflows, including using it as a proxy for model requests. Its monitoring, request controls, and testing tools support ongoing analysis and optimization. Helicone offers cloud-based service and support for self-hosted deployment. In its March 3, 2026 acquisition announcement, Helicone said its services would remain live for the foreseeable future in maintenance mode, with security updates, new models, and bug and performance fixes continuing. Consider that maintenance status when planning a long-term dependency. ## O que Helicone pode fazer? ### Usage, cost, and latency monitoring Track model usage, spending, and response latency to identify expensive workloads and performance bottlenecks. ### Gateway request controls Route model requests through a gateway with caching, rate limits, and automatic fallbacks. Model-swapping capabilities support changes to the models used by an application. ### Custom property analytics Attach custom properties to requests for more detailed analysis of application activity. Sessions and user analytics provide additional ways to examine usage. ### Queries, alerts, and reports Pro includes Helicone Query Language (HQL), alerts, and reports for investigating recorded activity and monitoring ongoing operations. ### Prompt and testing tools Use the playground to test prompts, sessions, and traces, and work with prompts, scores, and datasets to inspect model behavior. ### Retention and data access Plan-specific retention ranges from seven days on Hobby to indefinite retention on Enterprise. The pricing comparison also lists API access and data export. ## Casos de uso ### Investigate slow AI responses When an application becomes slow, developers can examine latency alongside recorded requests and session activity to locate the workloads that need attention. ### Analyze spending by application context Use custom properties to distinguish workloads and examine their model usage and costs, helping teams decide where optimization efforts should focus. ### Manage repeated requests and failures Use gateway caching for repeated requests and configure rate limits and automatic fallbacks as part of an application's request-handling workflow. ### Support organizational deployment requirements Teams that need control over infrastructure can explore self-hosting. Enterprise packages include on-premises deployment, SAML SSO, and a custom MSA. ## Perguntas frequentes ### What does the free plan include? Hobby includes 10,000 requests per month, 1 GB of storage, one seat, and one organization. Its listed data retention is seven days. This is an observability allowance; model inference through the gateway requires credits and should be budgeted separately. ### How are paid plans priced? Pro costs $79 per month and Team costs $799 per month, with usage-based charges applying to both. Both offer a seven-day free trial. Enterprise pricing is available through sales. ### How can I integrate Helicone? Helicone supports integration into existing AI workflows, including a gateway proxy for AI model requests. Its documentation provides a quick-start guide, gateway integration instructions, custom property guidance, and authentication details. ### Which plans support larger teams? Pro includes unlimited seats and one organization. Team adds five organizations and a dedicated Slack channel. Enterprise includes unlimited organizations, SAML SSO, on-premises deployment, and a custom MSA. ## Informações rápidas - **Category**: AI gateway and LLM observability - **Primary users**: Developers and teams building AI applications - **Deployment**: Cloud-based and self-hosted - **Source code**: Open source; github.com/Helicone/helicone - **Official website**: https://helicone.ai