Helicone

Helicone acts as a transparent HTTP/HTTPS proxy that intercepts all outbound LLM API calls from applications to external providers (OpenAI, Anthropic, etc.) without requiring code changes. Requests are routed through Helicone's gateway infrastructure, logged, and forwarded to the target provider with response data captured for observability. The proxy pattern enables one-line integration by replacing provider API endpoints with Helicone's proxy URL, maintaining full API compatibility while capturing request/response metadata.

Helicone automatically captures and stores all LLM API request/response pairs with extracted metadata including model name, token counts, latency, cost, user identifiers, and custom properties. Logs are persisted in a queryable database with configurable retention periods (7 days free tier to forever on enterprise). The logging system operates asynchronously to minimize impact on application latency and supports batch ingestion at rates from 10 logs/min (hobby) to 30,000 logs/min (enterprise).

Helicone's Playground is an interactive web interface for testing LLM prompts and models in real-time. Users can write prompts, select models, adjust parameters (temperature, max tokens, etc.), and execute requests against live LLM providers. The Playground supports testing against datasets and comparing outputs across models or prompt versions. Results are displayed with metadata (latency, cost, tokens) and can be saved for later reference.

Helicone's proxy gateway abstracts away provider-specific API differences, enabling applications to switch between LLM providers (OpenAI, Anthropic, Cohere, etc.) with minimal code changes. The gateway translates requests to provider-specific formats and normalizes responses, exposing a unified interface. Provider selection can be configured per request or globally, with fallback logic for provider failures. This abstraction enables cost optimization and redundancy without application-level provider handling.

Helicone exposes a REST API for programmatic access to logs, analytics, and configuration. The API supports querying request logs, retrieving cost data, managing prompts, and configuring alerts. Rate limits are tiered by subscription level (10 calls/min hobby, 1,000 calls/min team). API authentication uses API keys with optional IP whitelisting. The API enables building custom dashboards, reports, and integrations without dashboard access.

Helicone offers on-premises deployment option (enterprise tier only) enabling organizations to run the entire observability platform within their own infrastructure. On-prem deployments provide data residency compliance, network isolation, and full control over retention and access. The deployment includes the proxy gateway, logging backend, dashboard, and API. Organizations maintain their own infrastructure and are responsible for scaling, backups, and updates.

Helicone automatically calculates LLM API costs per request based on provider pricing (tokens × rate) and aggregates costs by user, session, or custom properties. Cost data is displayed in the dashboard with breakdowns by model, provider, and time period. The system supports custom user identifiers and session tracking to enable cost attribution and chargeback analysis. Cost calculations are performed server-side using current provider pricing rates.

Helicone's caching layer intercepts LLM requests at the proxy level and stores responses in a distributed cache, returning cached results for identical or semantically similar requests without calling the LLM provider. The cache supports configurable TTL and eviction policies, with cache hits/misses tracked in logs. Caching works transparently across all LLM providers by matching request payloads (model, prompt, parameters) and returning stored responses, reducing API costs and latency for repeated queries.

Helicone enforces rate limits at the gateway level, throttling requests based on configurable per-user, per-model, or global limits. When rate limits are exceeded, the system can automatically fall back to alternative models or providers (e.g., GPT-4 → GPT-3.5-turbo) to maintain service availability. Rate limit policies are configured in the dashboard and applied uniformly across all application instances without code changes. Fallback logic is defined as rules mapping primary models to alternatives.

Helicone tracks user sessions and interactions across multiple LLM requests, aggregating metrics like session duration, request count, cost per session, and user engagement patterns. Custom properties can be attached to requests to enable segmentation by feature, cohort, or experiment. Analytics are visualized in the dashboard with filters and breakdowns by user, time period, and custom dimensions. Session tracking requires explicit user identifiers in request headers or metadata.

HQL is a custom query language (Pro+ tier) enabling developers to write complex queries against the request log database to extract, filter, and aggregate data. HQL supports filtering by request properties (model, user, cost, latency), aggregation functions (sum, avg, count), and time-based grouping. Queries are executed server-side and results returned as structured data. HQL abstracts away the underlying database schema, providing a domain-specific interface for LLM observability queries.

Helicone sends webhook notifications for configurable events (request completion, cost threshold exceeded, error occurred, etc.) to external systems. Webhooks are HTTP POST requests containing event metadata and can trigger downstream workflows in Slack, PagerDuty, or custom applications. Webhook configuration includes event filtering, retry logic, and payload customization. Webhooks enable real-time alerting and integration with external monitoring/incident management systems.

Helicone's Prompts feature enables storing, versioning, and managing LLM prompts in a centralized registry. Prompts can be tagged, versioned, and deployed to production with rollback capabilities. The system tracks which prompt version was used for each request, enabling analysis of prompt performance and A/B testing. Prompts are accessed via API or dashboard, with version history and metadata stored in Helicone's database.

Helicone's Datasets feature enables creating curated datasets of LLM inputs/outputs for evaluation and testing. Datasets can be created from production logs or manually uploaded, with support for custom evaluation metrics and scoring. The Scores feature allows attaching evaluation scores (e.g., correctness, relevance) to requests, enabling quality tracking over time. Datasets and scores are used for prompt testing and model evaluation in the Playground.