Langfuse
Coding
Open-source LLM observability and evaluation
llm-observabilityopen-sourcetracing
Free planView →
Helicone
Open-source LLM observability and AI gateway in one line of code.
Free plan Claimed API Self-hosted Teams
Toolglade’s take
Helicone is one of the easiest ways to add LLM logging and a gateway, and the open-source, self-hostable version is genuinely appealing and free. The critical caveat is that after the March 2026 Mintlify acquisition, the product is in maintenance mode with no roadmap or new features. For a self-hosted logging layer it can still work well, but weigh the frozen roadmap and possible migration before making a new long-term production commitment.
About Helicone
Helicone is an open-source LLM observability platform and AI gateway that adds logging, tracing, caching, rate limiting, and provider failover with a one-line base-URL change. It is Apache 2.0 licensed and free to self-host with a lightweight Docker image. Since its March 2026 acquisition by Mintlify, the product operates in maintenance mode with no new features, which is an important consideration for new adopters.
Helicone became popular as a dead-simple way to add observability to LLM applications: developers change their base URL, and Helicone proxies traffic, logging every request as a trace while adding caching, rate limiting, and provider failover. It also ships a Rust-based AI Gateway that handles routing, load balancing, and caching across many providers. The platform is Apache 2.0 licensed and free to self-host, with a tiny Docker image that runs on Kubernetes, AWS ECS, Google Cloud Run, or bare metal. A hosted service offers a generous free tier plus paid Pro, Team, and Enterprise plans with usage-based overages. An important status change: Helicone was acquired by Mintlify in March 2026 and now operates in maintenance mode, receiving security patches, bug fixes, and new-model support but no new features or roadmap. The hosted service remains live for the foreseeable future, but this materially changes its suitability for new long-term commitments, and some users are being advised to migrate.
TL;DR
Helicone is an open-source LLM observability platform and AI gateway that adds logging, caching, rate limiting, and provider failover with a one-line integration. It is Apache 2.0 licensed and free to self-host, with a generous hosted free tier and paid plans from about $79 per month. The most important fact for 2026 is that Helicone was acquired by Mintlify and now operates in maintenance mode with no new features, which affects its suitability for new long-term adoption.
Company overview
Helicone was founded by Justin Torre and Cole Gottdank and went through Y Combinator's W23 batch. It grew quickly as a lightweight observability layer, reporting roughly 16,000 organizations and trillions of tokens processed by the time of its acquisition.
In March 2026, Helicone was acquired by Mintlify. Terms were not disclosed, and the product moved into maintenance mode as a result.
Product features
Helicone provides request logging and tracing, sessions, user analytics, dashboards, alerts, reports, and a queryable log language called HQL. Its Rust-based AI Gateway adds routing and load balancing, response caching including semantic caching, rate limiting, and automatic provider failover.
It integrates with major providers including OpenAI, Anthropic, Azure, Gemini, AWS Bedrock, LiteLLM, Together AI, and OpenRouter, and can be fully self-hosted via a small Docker image.
Target market
AI engineering teams that want quick, inexpensive LLM observability and a gateway, including teams that prefer to self-host the open-source version for data control.
Buyer personas
End users
AI and backend engineers who instrument LLM applications for logging, cost tracking, and reliability.
Buyers
Engineering managers and technical leads choosing an observability and gateway layer.
Key influencers
Platform engineers, open-source advocates, and cost-conscious developers.
Ideal customer profile
A developer or small AI team that wants fast, cheap LLM logging and gateway features, comfortable self-hosting the open-source edition and aware of the maintenance-mode status.
Funding & performance
Helicone was part of Y Combinator W23 and disclosed roughly $500K in seed funding. No large priced venture round was publicly disclosed before its acquisition by Mintlify in March 2026, terms of which were not disclosed.
Pros & cons
Pros
- Extremely easy integration via a one-line base-URL swap
- Open source under Apache 2.0 and free to self-host
- Tiny Docker image that runs almost anywhere
- Gateway features like caching, failover, and rate limiting cut cost and improve reliability
- Generous free tier of 10,000 requests per month
- Broad provider coverage
- HQL query language for flexible log analysis
Cons
- Product is in maintenance mode after the Mintlify acquisition, with no roadmap
- Long-term hosted-service viability is uncertain and migration may be needed
- Usage-based overages mean the Pro price is a floor, not a cap
- Short retention on lower tiers (7 days free, 30 days Pro)
- Large price jump from Pro to Team for compliance features
- Evaluation and experimentation depth is lighter than eval-first tools
- The proxy model adds a hop and a dependency in the request path
Pricing plans
Hobby
$0
- 10,000 requests/mo
- 1GB storage
- 1 seat
- 7-day retention
Pro
$79 / month
- Unlimited seats
- Alerts and reports
- HQL query language
- 1-month retention
- Usage-based overages
Team
$799 / month
- SOC 2 and HIPAA compliance
- 5 organizations
- All Pro features
- Priority support
Enterprise
Custom
- SAML SSO
- Custom MSA
- On-prem deployment
- Dedicated support
Key features
API
Team collaboration
Self-hosted
Integrations
OpenAI, Anthropic, Azure OpenAI, Google Gemini, AWS Bedrock, LiteLLM, Together AI, OpenRouter
Input types
text
Output types
text
Best For
quick LLM request logging, self-hosted observability, gateway caching and failover, cost and usage monitoring
Compare key features
View all alternatives →Feature
Helicone
Langfuse
Portkey
Pricing
Freemium
Freemium
Freemium
Free plan
Yes
Yes
Yes
Free trial
No
Yes
No
API
Yes
Yes
Yes
Self-hosted
Yes
Yes
Yes
Team support
Yes
Yes
Yes
Frequently asked questions
Is Helicone still being developed?+
After the March 2026 Mintlify acquisition, Helicone is in maintenance mode. It receives security patches, bug fixes, and new-model support, but no new features or roadmap. The hosted service remains live for the foreseeable future.
Is Helicone open source?+
Yes. Helicone is licensed under Apache 2.0 and is free to self-host, including its Rust-based AI Gateway. The Docker image is small and runs on Kubernetes, ECS, Cloud Run, or bare metal.
How does the integration work?+
In most cases you change your LLM provider base URL to route requests through Helicone and add a Helicone key. That single change enables logging, tracing, and optional gateway features.
What providers does Helicone support?+
It supports OpenAI, Anthropic, Azure OpenAI, Google Gemini, AWS Bedrock, LiteLLM, Together AI, OpenRouter, and more.
Should I adopt Helicone for a new project?+
For a quick, cheap self-hosted logging layer it can still work well, but the frozen roadmap after the Mintlify acquisition is a real risk for long-term production use. Consider migration paths before committing.
Reviews (0)
Loading reviews…
Compare
Compare Helicone with other AI tools
Side-by-side pages for pricing, features, and best-fit use cases.
From the blog
All articles →Similar tools you may like
Free Plan
Portkey
Coding
Open-source AI gateway and control plane for production Gen AI.
ai-gatewayllmopsopen-source
Free planView →
Hugging Face
Coding
The open hub for machine learning models, datasets, and demos.
open-sourcemachine-learningmodel-hub
Free planView →
Groq
Coding
Very fast LLM inference on custom LPU hardware.
inferencellm-apilow-latency
Free planView →