# Helicone/helicone

🧊 Open source LLM observability platform. One line of code to monitor, evaluate, and experiment. YC W23 🍓

Repository: https://github.com/Helicone/helicone
Canonical: https://ross.abutalabs.com/products/helicone
Homepage: https://www.helicone.ai
Language: TypeScript
License: Apache-2.0
License Family: permissive
Topics: large-language-models, prompt-engineering, agent-monitoring, analytics, evaluation, gpt, langchain, llama-index, llm, llm-cost, llm-evaluation, llm-observability, llmops, monitoring, open-source, openai, playground, prompt-management, ycombinator
Last push: 2026-08-26T00:04:40+00:00

## Health v2 (maintenance only)
Score: 79/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 44, longevity 93
- inputs: {"age_days": 1310, "days_push": 8, "days_rel": 377, "gap_med": 1, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 6104, forks 663 (observed 2026-08-28T04:09:35.072200+00:00)

## What it is
Helicone is an open-source LLM observability platform and AI gateway that logs, monitors, and analyzes LLM API requests with a one-line integration. It provides an OpenAI-compatible gateway to 100+ models with routing, fallbacks, caching, cost/latency tracking, agent tracing, and prompt management.

## Use cases
- monitor llm api costs and latency
- log openai and anthropic requests
- trace ai agent calls
- route llm requests across providers with fallbacks
- version and test prompts
- evaluate llm output quality
- track token usage per user
- debug chatbot and rag pipelines

## When to choose
- you need production observability for LLM calls with minimal code changes
- you want a unified gateway across OpenAI, Anthropic, Google, and other providers
- you need cost, latency, and quality dashboards for AI apps
- you want prompt versioning, datasets, and a playground for iteration
- you want a self-hostable open-source option with a hosted tier

## When to avoid
- you only need generic APM without LLM-specific features
- you require strict data residency but cannot self-host and cannot use cloud logging
- your stack is entirely offline or local models with no API calls to observe

## Facets
- artifact type: service
- maturity: active
- function: monitoring, tracing, analytics, llm-inference, prompt-engineering, api-gateway, caching, rate-limiting, benchmarking, webhook
- domain: large-language-models, artificial-intelligence, monitoring, developer-tools, analytics
- platform: self-hosted, cloud, python, cross-platform
- tags: llm-observability, ai-gateway, llmops, llm-cost-tracking, prompt-management, agent-tracing, openai-proxy, model-routing, ai-agents, web-server, nodejs

## Member repositories
- Helicone/helicone (main) score 79

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:35.072200+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:48:14.926366+00:00, confidence not recorded.
  - readme: https://github.com/Helicone/helicone (fetched 2026-08-28T04:09:35.072200+00:00, sha b2311d4e7ece)
  - homepage: https://www.helicone.ai (fetched 2026-08-29T08:45:07.414405+00:00, sha 3fb16e125f1b)
  - site_page: https://docs.helicone.ai (fetched 2026-08-29T08:45:07.418571+00:00, sha e6fb279cf198)
  - site_page: https://docs.helicone.ai/integrations/openai/javascript (fetched 2026-08-29T08:45:07.422766+00:00, sha 09d2d2df402b)
  - site_page: https://docs.helicone.ai/integrations/anthropic/javascript (fetched 2026-08-29T08:45:07.424302+00:00, sha 02307ac08472)
  - site_page: https://docs.helicone.ai/integrations/azure/javascript (fetched 2026-08-29T08:45:07.426158+00:00, sha b9682e095b76)
  - site_page: https://docs.helicone.ai/getting-started/integration-method/litellm (fetched 2026-08-29T08:45:07.427894+00:00, sha 39b09363d6bb)
  - site_page: https://docs.helicone.ai/getting-started/integration-method/anyscale (fetched 2026-08-29T08:45:07.429677+00:00, sha 5b63dc6d41fe)
  - site_page: https://docs.helicone.ai/getting-started/integration-method/together (fetched 2026-08-29T08:45:07.472408+00:00, sha c965832cb0d3)
  - registry_npm: https://registry.npmjs.org/helicone (fetched 2026-08-29T08:45:07.486441+00:00, sha 00242d678f47)
- Data as of 2026-08-30T08:39:29.467469+00:00.