Read our 2026 review of Helicone: open-source LLM observability and AI gateway features, pricing tiers, the Mintlify acquisition, pros, cons and alternatives.
Category
AI Infrastructure & MLOps
Pricing
Freemium (usage-based paid tiers), from Free (Hobby plan, 10,000 requests/month)
Helicone is an open-source LLM observability platform and AI gateway founded in 2023 by Justin Torre and Cole Gottdank, and a graduate of Y Combinator's Winter 2023 batch. It gives developers a unified way to log, monitor, evaluate and optimize the cost and performance of applications built on OpenAI, Anthropic and other LLM providers, with source code publicly available on GitHub.
In March 2026, Helicone was acquired by Mintlify, the AI-era documentation and knowledge-infrastructure company, after growing to process 14.2 trillion tokens for more than 16,000 organizations as an independent company. Helicone now operates in maintenance mode, continuing security updates and new model support under Mintlify.
Key Features
Helicone functions as both a passive observability layer and an active AI gateway: it logs every prompt, completion, latency figure and calculated cost across LLM providers, and it can route and fall back traffic across multiple providers or models. Advanced users get access to HQL, Helicone's dedicated query language for analyzing logged AI traffic.
The platform also includes evaluation tooling to test and score model outputs, alerting for cost spikes or anomalies, and compliance features like SOC-2 and HIPAA support on higher tiers.
Pricing
Helicone's free Hobby tier includes 10,000 requests a month, 1 GB of storage and a single seat, with 7-day data retention. Paid tiers scale from Pro at 79 dollars a month (unlimited seats, HQL, 1-month retention) to Team at 799 dollars a month (SOC-2 and HIPAA, 3-month retention) up to custom-priced Enterprise with SAML SSO, on-premises deployment and forever retention.
Discounts are available for early-stage startups (50 percent off the first year), non-profits, open-source projects (100-dollar annual credit) and students.
Key Features
One-line-of-code LLM logging — A lightweight proxy or SDK integration captures every prompt, completion, latency and cost figure across LLM providers with minimal setup.
AI gateway with multi-provider routing — Traffic can be automatically routed or failed over across multiple LLM providers and models to maintain uptime and manage cost.
HQL (Helicone Query Language) — A dedicated query language, available on paid tiers, for advanced analysis of logged AI request and cost data.
Cost and token tracking — Per-request cost is automatically calculated across models and providers, giving teams visibility into LLM spend without manual tracking.
Evaluation tooling — Built-in tools for testing and scoring model outputs help teams catch regressions in quality before they reach production.
Alerting on anomalies and cost spikes — Configurable alerts notify teams of unusual latency, error rates or cost spikes across their LLM traffic.
SOC-2 and HIPAA compliance — Available on the Team and Enterprise tiers for organizations with regulatory or security compliance requirements.
Open-source, self-hostable codebase — The full platform is open source on GitHub, giving developers transparency and the option to self-host rather than rely solely on Helicone's cloud offering.
Pros & Cons
Pros
Genuinely open source, with full platform code available on GitHub
Generous free Hobby tier covering 10,000 requests a month with no cost
Functions as both an observability platform and an AI gateway with routing and fallback
SOC-2 and HIPAA compliance available for regulated and enterprise customers
Backed by Y Combinator credibility and now Mintlify's resources post-acquisition
Cons
Now in maintenance mode following the Mintlify acquisition, meaning independent roadmap development has slowed
Advanced querying via HQL and unlimited seats require upgrading to the Pro tier or above
Data retention on lower tiers is short (7 days on Hobby, 1 month on Pro), which may not suit teams needing longer audit trails
Team and Enterprise pricing is a significant jump from the Pro tier for compliance-focused organizations
As with any proxy-based observability tool, routing LLM traffic through Helicone adds a dependency in the request path
Pricing
Hobby Free N/A
Pro $79/month Monthly
Team $799/month Monthly
Enterprise Custom Contact for pricing
Frequently Asked Questions
What does Helicone do?
Helicone is an LLM observability platform and AI gateway that logs, monitors and helps optimize the cost and performance of applications built on LLM providers like OpenAI and Anthropic, with a single line of code.
Is Helicone free?
Yes, Helicone's Hobby tier is free and includes 10,000 requests a month, 1 GB of log storage and 7-day data retention. Paid plans start at 79 dollars a month for the Pro tier.
Is Helicone open source?
Yes, Helicone's full platform code is open source and publicly available on GitHub, which also allows for self-hosting.
Who founded Helicone?
Helicone was founded in 2023 by Justin Torre and Cole Gottdank, and it went through Y Combinator's Winter 2023 (W23) batch.
Was Helicone acquired?
Yes, Helicone was acquired by Mintlify, the AI-era documentation and knowledge infrastructure company, in a deal announced in March 2026. Helicone now operates in maintenance mode under Mintlify.
Does Helicone support routing between multiple LLM providers?
Yes, Helicone functions as an AI gateway that can route traffic and handle fallback across multiple LLM providers and models, in addition to its observability and logging features.