Skip to content
agentscamp
Tool

Helicone

Open-source LLM observability and AI gateway with one-line integration — logging, tracing, caching, and cost/latency tracking across providers.

open sourceobservability
Updated Jun 4, 2026
observabilitygatewaycachingcost-controlopen-source
Visit website Source
View as Markdown

Helicone is an open-source LLM observability platform and AI gateway with a one-line integration — logging, tracing, caching, and cost/latency tracking across providers. Note: Mintlify acquired Helicone in March 2026 and it's now in maintenance mode (security and bug fixes only, no new features), though the Apache-2.0 proxy still works and is self-hostable.

Helicone is an open-source LLM observability platform with a built-in AI gateway, known for a famously low-friction setup: change your base URL (or add a header) and your calls are logged, traced, and analyzed — no SDK rewrite. On top of monitoring it offers caching and rate limiting at the proxy, cost and latency tracking, prompt management, and datasets/evals.

For a cost-and-latency stack its draw is the one-line on-ramp to per-call cost and latency visibility — you can't optimize what you can't see — plus proxy-level caching to cut spend on repeated calls.

Highlights

  • One-line integration — proxy via a base-URL change, or async logging; no rewrite of your call sites.
  • Observability — requests, sessions, and traces with cost and latency per call, user, and model.
  • Caching & rate limiting — at the proxy, to cut the cost and latency of repeated calls.
  • Prompt management, datasets & evals — version prompts and score traffic.
  • Open source & self-hostable — Apache-2.0, with Docker/Helm for self-hosting.

In an AI-assisted workflow

# point the OpenAI client at Helicone's proxy — one line, then traffic is observable
from openai import OpenAI
client = OpenAI(base_url="https://oai.helicone.ai/v1",
                default_headers={"Helicone-Auth": f"Bearer {HELICONE_API_KEY}"})

WARNING

Status (2026): Mintlify acquired Helicone in March 2026, and the product is now in maintenance mode — security patches, bug fixes, and new-model support continue, but there are no new features or roadmap, and Mintlify is helping customers migrate to other platforms. The open-source proxy still works and the Docker image is current, so existing self-hosted deployments keep running; new projects should weigh that it is no longer actively developed.

Good to know

Helicone is open source (Apache-2.0) and free to self-host; a hosted cloud with a free tier and paid plans is also available. It was used in production by 16,000+ organizations at the time of the acquisition. Given the maintenance-mode status above, teams starting fresh should compare actively-developed observability platforms like Langfuse and LangSmith, and the gateway-first Portkey — see LLM Gateways Compared.

Related