Helicone FAQs

Is there an impact to the latency of the calls to LLM?

Helicone Proxies your requests through globally distributed nodes running on Cloudfare Workers. This means that the latency is minimal and the requests are routed to the closest server to the end user.

I don't want to use Helicon's Proxy, can I still use Helicone?

Yes, you can use still use Helicone to log your requests using the Helicone SDK's Async Integration without proxying.

What makes Helicone different from other LLM observability tools?

Helicone is unique in its open-source nature, providing developers with complete control over their data and the ability to customize its features. Helicone's focus on production-level performance and its wide range of integration options set it apart from other tools in the market.

What are the pricing plans for Helicone?

Helicone offers a free plan for individual developers and teams getting started. For more advanced features and higher usage, Helicone offers paid plans with different tiers based on your specific needs.

How can I get started with Helicone?

Getting started with Helicone is easy. You can sign up for a free account on the Helicone website and start logging your LLM requests immediately. Helicone provides comprehensive documentation and a helpful community forum to guide you through the process.

• Begin by creating a Helicone account and integrating your preferred LLM provider (e.g., OpenAI, Anthropic). This establishes a connection to monitor your application's performance.
• Utilize the Helicone dashboard to monitor requests, sessions, and user interactions in real-time. This provides a comprehensive overview of your application's activity.
• Employ Helicone's logging functionality to pinpoint errors and debug your LLM application. Detailed traces simplify troubleshooting and root cause analysis.
• Leverage the evaluation tools to assess LLM performance and identify areas for improvement. Features like LLM-as-a-judge facilitate objective quality assessment.
• Use the Experiments feature to test different prompts and track their effects on performance. This allows data-driven prompt optimization without modifying core code.
• Review the collected data and metrics to gain insights into your LLM application's performance and user behavior. The results inform iterative development and optimization.
• Explore advanced features like webhooks and user metrics for deeper insights and automated alerts. This streamlines monitoring and proactive issue detection.