Helicone
Helicone builds open-source infrastructure for teams developing AI applications: a unified platform to route, observe, debug, monitor and optimize LLM applications across providers. Its AI Gateway provides one OpenAI-compatible interface to 100+ models with routing, fallbacks and observability.
At a Glance
- AI-native startups
- Software developers and data scientists
- Enterprise AI teams
- Financial services
- +4 more
AI Tools by Helicone
(2)
Helicone AI Gateway
Open Source LLM Gateway API
Helicone
LLM Observability and Analytics
Discussions
No discussions yet
Be the first to start a discussion about Helicone
Latest News
Helicone is joining Mintlify; Helicone says it was acquired and its team is joining Mintlify in San Francisco.
Published MCPs: What, Why, and How, covering Model Context Protocol and its use for connecting LLMs to tools and data.
Launched the cloud-hosted Helicone AI Gateway with passthrough billing, 100+ models, automatic fallbacks and 0% markup.
Introduced the Helicone AI Gateway for unified multi-provider access and reliability.
Products & Services
Open-source dashboard and proxy/SDK platform for logging LLM requests, tracing sessions, analyzing cost, latency, users and prompts, and debugging AI applications.
OpenAI-compatible gateway for 100+ models and providers, with intelligent provider/model routing, failover, usage-based controls, caching and built-in observability; available as hosted cloud and self-hosted open source.
Self-hosted deployment of Helicone for organizations requiring data sovereignty, compliance and control over infrastructure; supports Docker and Kubernetes/Helm deployment.
Version, deploy and iterate prompts through the platform, with prompt testing and playground workflows.
Market Position
Helicone positions itself as an open-source, developer-first alternative to closed or more narrowly scoped LLM observability tools, combining request observability with an AI Gateway, multi-provider routing, cost controls and self-hosting. Its own comparison materials identify LangSmith, Langfuse, Galileo and OpenRouter as relevant alternatives or adjacent competitors; Helicone differentiates with open source, provider flexibility, integrated routing and 0% gateway markup.
Leadership
Founders
Justin Torre
Founder/CEO; previously a developer evangelist and teacher at Apple, with additional experience at Intel and Sisu Data.
Barak Oshri
Co-founder; previously a machine-learning engineer at Sisu Data, and earlier a researcher at the Stanford AI Lab and teaching assistant at Stanford.
Scott Nguyen
Co-founder; brought UX and finance experience from Tesla, Bain Capital and DraftKings.
Cole Gottdank
Co-founder and go-to-market/product contributor; he later led Helicone's AI Gateway launch and joined Mintlify after the acquisition.
Executive Team
Justin Torre
Founder and CEO (prior to acquisition)
Developer evangelist and teacher at Apple; also worked across Intel and Sisu Data before founding Helicone.
Cole Gottdank
Co-founder (prior to acquisition)
Helicone co-founder and public-facing product/GTM contributor; authored the AI Gateway launch announcement and joined Mintlify after the acquisition.
Board of Directors
Founding Story
Helicone was started during Y Combinator's Winter 2023 batch when the founders saw that the first wave of GPT-wrapper and AI-native products lacked an observability layer. They set out to make it straightforward for developers to observe and manage language-model usage, initially with an open-source proxy that logged requests, exposed cost/performance analytics, cached completions and handled retries/routing.
Business Model
Revenue Model
Usage-based SaaS pricing for hosted observability, with paid subscriptions and usage charges; the AI Gateway uses passthrough provider billing with 0% markup. Enterprise plans add custom contracts, compliance, SSO and on-premises deployment.
Pricing Tiers
10,000 requests per month, 1 GB storage, 1 seat and 1 organization.
Unlimited seats, alerts and reports, HQL; usage-based pricing also applies.
Five organizations, SOC 2 and HIPAA compliance, dedicated Slack channel; usage-based pricing also applies.
Custom MSA, SAML SSO, on-prem deployment and bulk cloud discounts.
Target Markets
- AI-native startups
- Software developers and data scientists
- Enterprise AI teams
- Financial services
- Healthcare
- Government
- Monitoring and debugging production LLM applications and AI agents
- Controlling and reducing model/API spend through caching, routing and cost analytics
- Building multi-model applications without maintaining separate provider integrations
- Reliability engineering with automatic failover, retries and rate-limit management
- Prompt experimentation, evaluation and optimization
- Self-hosted AI observability for privacy-sensitive and regulated organizations
- Together AI
- QA Wolf
- Clay
- Singapore Airlines