Looking for an alternative to Cloudflare AI Gateway?
Cloudflare AI Gateway is a free control layer in front of model providers, paid through your own keys or Unified Billing credits; Standard Compute routes Claude Code, Codex and OpenAI-compatible agents under one flat monthly plan per key.
TL;DR
Stay on Cloudflare if your apps already run there and you bring your own keys. Switch if you want Claude Code and Codex routed across models on a fixed monthly bill.
What Cloudflare AI Gateway offers
Core features are free on every Cloudflare plan. Unified Billing adds a 5% fee on credit purchases ($100 of credits costs $105), with provider inference passed through at no markup. Logs for gateways created from September 24, 2026 follow Workers Logs pricing. (Verified 2026-10-13.)
- Free gateway with caching, analytics and rate limiting on any Cloudflare account
- Spend limits scoped by model, provider or metadata such as user or team, that can block or fall back to a cheaper model
- Auto Router (cloudflare/auto) classifies each request's type and difficulty and weighs quality against cost
- Coding-agent guides for Claude Code, Codex, OpenCode, Pi, Gemini CLI and more
- Sits next to Cloudflare's security stack, with DLP free on all plans
Why people look for an alternative
- Unified Billing traffic is rate-limited per gateway; only traffic on your own provider keys is exempt
- Auto Router accepts Chat Completions and Responses requests, not Anthropic Messages, so Claude Code traffic isn't auto-routed; Dynamic Routing also takes the Chat Completions format only
- Codex through the gateway works with OpenAI models only, by Cloudflare's own docs
- Spend limits are "eventually consistent": cost is recorded after a request completes, so bursts can pass the limit
- AI Gateway is not covered by Cloudflare's Data Localization Suite regional services
Which one fits your work
| Pick | When |
|---|---|
| Standard Compute | Claude Code on the Anthropic Messages API and Codex on the Responses API, both routed across Claude, GPT and open models |
| Compute included in the plan, with a published budget, instead of credits plus a fee | |
| Europe or US region rules per key, team or company, applied to the provider endpoints that serve your requests | |
| Company and agency workspaces with a plan or limit per developer key | |
| Cloudflare AI Gateway | Your apps already run on Cloudflare and you want the gateway's caching, analytics and DLP for free |
| You bring your own provider keys and want a gateway that adds no fee at all | |
| You want to build your own routing flows with conditions, percentage splits and budget fallbacks | |
| Pay-as-you-go use that is light or uneven |
Switching takes one config change
Standard Compute is OpenAI-compatible, so any tool or SDK that lets you set a custom base URL migrates in minutes:
Base URL = https://api.stdcmpt.com/v1 API key = your Standard Compute key Model = standardcompute
Setup guides for OpenClaw, Hermes, OpenCode, Cursor, Cline, Aider and more are on the integrations page.
FAQ
Is Cloudflare AI Gateway free?
The gateway's core features are free on all Cloudflare plans. You pay for inference, either to your own providers or as Unified Billing credits with a fee on each purchase, and logs on new gateways follow Workers Logs pricing.
Can I use Cloudflare AI Gateway with Claude Code?
Yes. Point ANTHROPIC_BASE_URL at your gateway's /anthropic path and add a cf-aig-authorization header, paying with Unified Billing credits or your own Anthropic key. On Standard Compute you set ANTHROPIC_BASE_URL to https://api.stdcmpt.com and use your Standard Compute key.
Does Cloudflare AI Gateway choose the model for me?
Its Auto Router does, when you send model cloudflare/auto in the Chat Completions or Responses format. It picks from a Cloudflare-managed pool of Anthropic, OpenAI and xAI models, and adds open models such as DeepSeek or GLM only if you allow-list them in a header. Standard Compute's router picks from the up to five models you select, on all three agent APIs.
What is the Unified Billing rate limit?
200 requests per 60 seconds per gateway, answered with HTTP 429 above that. Requests that use your own provider keys don't count toward it.
Sources: AI Gateway pricing; Unified Billing; AI Gateway limits; Auto Router; Spend limits; Codex setup
