ngrok AI Gateway and Cloudflare AI Gateway side by side
| ngrok AI Gateway | Cloudflare AI Gateway | |
|---|---|---|
| What it is | One endpoint for hosted providers and self-hosted models | A gateway in front of model providers, run by Cloudflare |
| Gateway fee | $0.05 per million tokens on every successful request, also with your own keys | Core features free on all plans; 5% on Unified Billing credits |
| Paying for models | Prepaid credits ($5 minimum, $500 maximum balance, expire after a year), or your own provider keys | Your own provider keys, or Unified Billing credits by card |
| Self-hosted models | Ollama, vLLM and other servers you run | Not stated on the pages checked |
| Failover | Next key or model on 429, 401/403, timeouts and 5xx | Not stated on the pages checked |
| Logging | Usage recorded; prompt storage not stated | Logs on by default and can include prompt and response; opt out per gateway or request |
| Retention | Not stated | Zero data retention only for Unified Billing with Cloudflare-managed credentials |
| Processing in Europe | Not stated | Regional Services not supported for AI Gateway; log jurisdiction restrictions not supported |
| Spend control | HTTP 402 at zero balance; a maximum spend per period in auto top-up | Spend-limit rules per gateway by model, provider or metadata |
| Access control | Per access key: allowed providers and models | Per gateway |
| Single sign-on | SSO add-on at $10 per user a month; SCIM and audit logs at $15 more | Dashboard SSO free on all plans |
Where they differ most
Fees
ngrok bills a processing fee of $0.05 per million tokens on every request, with your own keys too. Cloudflare's core gateway features are free, and Unified Billing adds 5% to credit purchases.
Your own models
ngrok routes to hosted providers and models you run yourself through one endpoint, and fails over to the next key or model on rate limits and errors.
Logs and Europe
Cloudflare logs prompts and responses by default unless you opt out, and AI Gateway does not support Regional Services. ngrok does not state where its gateway processes requests.
Choose ngrok AI Gateway or Cloudflare AI Gateway
Choose ngrok AI Gateway if
- You serve some models yourself and want them behind the same endpoint.
- Failover across keys and models matters more than a free gateway.
- Your team already uses ngrok for endpoints and traffic.
Choose Cloudflare AI Gateway if
- Your company runs on Cloudflare and wants no gateway fee.
- Spend-limit rules by model, provider or user cover your cost control.
- You have no requirement to keep gateway logs in a specific jurisdiction.
A third option for companies
Both are gateways in front of model contracts or credits you manage. If you need Claude, GPT, Gemini and open models with company controls and processing in Europe, Standard Compute is a managed gateway: set a Europe rule for a company, team or key, and every request is served only by approved European inference endpoints.
Cost per step
Each step goes to the cheapest model that can do it and the cheapest provider fast enough to serve it.
Models
Claude, GPT and Gemini alongside open models such as GLM, DeepSeek, MiniMax and Kimi, through one API.
Processing in Europe
A Europe rule per company, team or key: every request, including retries and fallbacks, is served only by approved European inference endpoints, or refused before your prompt is sent.
Zero data retention
Required per company, team or key, and Standard Compute never trains on your prompts or outputs.
Single sign-on
Microsoft Entra ID or any SAML provider, and Claude Desktop and Claude Code sign-in with a Microsoft work account.
Budgets
A fixed monthly budget per person and app that never overruns, and spending caps per key and for the whole company.
Allow-lists
Approved providers per company, team or key, and approved models per key. Fallbacks never route around them.
Audit log
Admin changes with before and after values, and a 12-month export of every request's model, provider and region rule.
Sign our Data Processing Agreement in the dashboard, and see every sub-processor with its purpose and location.
Create a company workspaceTalk to usor write to contact@standardcompute.com
Questions
How much does ngrok AI Gateway cost?
ngrok charges a processing fee of $0.05 per million tokens on each successful request, paid from prepaid credits, including when you use your own provider keys. Model usage on ngrok-supplied keys is paid from the same credits.
Is Cloudflare AI Gateway free?
Cloudflare says core features are free on all plans. Credits bought through Unified Billing carry a 5% fee.
Which one logs prompts?
Cloudflare logs are on by default and can include prompts and responses; you can opt out per gateway or per request. ngrok's AI Gateway docs say usage is recorded and do not state whether prompts are stored.
Can either keep processing in Europe?
Cloudflare says Regional Services is not supported for AI Gateway. ngrok's AI Gateway docs do not state a region.
Is there a gateway with a Europe rule and zero data retention?
Turn on zero data retention in Standard Compute for your company, a team or a key, and every request is routed only to inference endpoints that commit to keeping no copy of your data. Add a Europe rule, and every request is served only by approved European inference endpoints.
Sources (checked 18 November 2026)
- ngrok AI Gateway overviewngrok.com
- ngrok AI Gateway creditsngrok.com
- ngrok AI Gateway: how it worksngrok.com
- ngrok access key configurationsngrok.com
- ngrok pricingngrok.com
- Cloudflare AI Gateway pricingdevelopers.cloudflare.com
- Cloudflare Unified Billingdevelopers.cloudflare.com
- Cloudflare AI Gateway loggingdevelopers.cloudflare.com
- Cloudflare Data Localization Suite compatibilitydevelopers.cloudflare.com
- Cloudflare dashboard SSOdevelopers.cloudflare.com