Buying snapshot
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Decision point | What Helicone states |
|---|---|
Licence | Apache v2.0 |
Deployment | Hosted service; self-host with Docker Compose; a production Helm chart for Enterprise workloads through enterprise@helicone.ai; on-prem deployment on Enterprise |
Integration | Change the baseURL to https://ai-gateway.helicone.ai and add a Helicone API key; async logging (OpenLLMetry) for JS/TS and Python |
Data retention | 7 days (Hobby), 1 month (Pro), 3 months (Team), Forever (Enterprise) |
Ingestion limit | 10 logs/min (Hobby), 1,000 (Pro), 15,000 (Team), 30,000 (Enterprise) |
Compliance | README: SOC 2 and GDPR compliant. SOC-2 and HIPAA compliance listed from the Team plan |
Our take
If your app already calls models through an OpenAI-compatible client and you want request logs and cost tracking without writing instrumentation, Helicone is a strong first pick, and the Apache 2.0 licence keeps self-hosting open. If you need on-prem with vendor support, budget for an Enterprise conversation from the start.
Evaluate Helicone
Where it stands out
Small first step. The quick start is a baseURL change plus an API key in your existing OpenAI client. Helicone calls it one line of code, and there is no tracing code to write before you see logs.
Gateway and observability in one product. The gateway reaches 100+ models with one API key, with intelligent routing and automatic fallbacks. The same requests feed traces, sessions and cost and latency tracking.
Prompts beyond logging. Helicone versions prompts using production data and deploys them through the AI Gateway without code changes. It also offers a playground and fine-tuning through partners OpenPipe or Autonomi.
Your data stays reachable. You can export metrics to PostHog in one line, and manage and export data with the Helicone API or an MCP server. Helicone also runs an open-source LLM Cost API with 300+ models and providers.
Product screenshot
Trade-offs to weigh
The gateway sits in your request path. The quick start sends traffic through ai-gateway.helicone.ai, so its availability matters to your app. If that is a problem, the async logging integration logs out of band, but it is a separate integration.
Self-hosting is several services. The stack includes a NextJS web app, a proxy worker, the Jawn log server, Supabase, ClickHouse and Minio. Manual deployment is not recommended, and the production Helm chart comes through the enterprise contact.
Ownership news is pending. Helicone's site says: "Helicone is no longer part of Mintlify. Details to come soon." Ask about the roadmap and contract terms before a long commitment.
A large step between Pro and Team. No plan sits between Pro and Team. Check whether you need SOC-2 and HIPAA compliance, more organizations or longer retention before you move up.
Pricing and plan conditions
Helicone bills a plan fee plus usage. Each paid plan includes 10K free requests and 1 GB free storage, then usage-based charges apply above that. Pro starts with a 7-day free trial. Self-hosting under the Apache 2.0 licence has no plan fee; you run the services yourself.
Scroll sideways to see all columns. With a keyboard, focus this table and use the arrow keys.
| Plan | Published price | What it adds |
|---|---|---|
Hobby | Free | 10,000 free requests, 1 GB storage, 1 seat, 1 organization |
Pro | $79 per month | Everything in Hobby, plus unlimited seats, alerts and reports, HQL (query language). Usage-based pricing applies |
Team | $799 per month | Everything in Pro, plus 5 organizations, SOC-2 and HIPAA compliance, a dedicated Slack channel. Usage-based pricing applies |
Enterprise | Contact us | Custom-built packages. Everything in Team, plus a custom MSA, SAML SSO, on-prem deployment, bulk cloud discounts |
Discounts on application: startups under 2 years old and $5M in funding get 50% off the first year. Open-source companies get a $100 credit for the first year. Non-profits get discounts depending on org size, and most students and educators use it free.
Public signals
About these signals
Verified factual signals from public sources. They indicate observable activity or interest, not total adoption, product quality, or cost.
14 GitHub commits 90d6.3k GitHub stars0 vulnerabilities across 2 packages
See all signals from 5 sources
npm · @helicone/helpers@1.8.3 · PyPI · helicone@1.0.14
Alternatives
Helicone's pricing page compares it with LangSmith, Braintrust, Arize AI and Langfuse. On LangSmith, Helicone's own short answer is that it "gives you more provider flexibility, is open-source, and scales more cost-effectively." That is the vendor's view, so run the same traces through each candidate before you choose.
Alternatives to Helicone
The reviewed substitutes for Helicone among the LLM observability & evaluation, and what would make each one the better answer.
Direct alternatives
Reviewed substitutes: products bought for the same job, where a team picks one.
- Arize Phoenix
- Both are open LLM observability tools a team runs itself. Helicone integrates as a proxy and Phoenix through OpenTelemetry, but the purchase decision is one or the other.
- Braintrust
- Both are LLM observability platforms with free tiers, bought for visibility into AI application behaviour and cost.
- Langfuse
- Both are open-source LLM observability platforms with a hosted tier, bought to see and cost AI requests.
Other approaches
A different approach to the same problem. Each substitutes only for the workload named beside it.
- LiteLLM
- Helicone's proxy and LiteLLM's gateway both sit in front of model providers, but Helicone is bought for observability and LiteLLM for routing and governance. They compete only for a small team that wants one service instead of two.Applies to: A single proxy providing both routing and request visibility