Is Conifer an alternative to Cloudflare AI Gateway?

Yes, for sending requests to models from many providers with fallbacks, spending limits and a cost for each request. Cloudflare AI Gateway also offers response caching, guardrails and data loss prevention, and its core features are free. Conifer sells model usage from a prepaid balance, serves a named model as named or refuses it, and states the exact cost of each request.

Can I use Cloudflare AI Gateway and Conifer together?

Yes. Cloudflare AI Gateway accepts custom providers at any HTTPS address, so you can add Conifer as a provider and keep Cloudflare’s caching, rate limits and logs in front of it.

Do I need provider accounts to use Conifer?

No. One Conifer key covers every model in the catalog, paid from a prepaid balance. You can also put your own Azure OpenAI resource behind the same key, and Conifer adds no fee on that usage.

Does Conifer store my prompts?

The gateway does not store the prompt of a standard request. It holds a response for ten minutes so that a retried request is not charged twice, and keeps request details such as the model, token counts and cost for billing.

Comparison

Conifer and Cloudflare AI Gateway compared

Both are hosted gateways that sit in front of AI models from many providers. This page sets out how Conifer and Cloudflare AI Gateway each work, using each company’s own published documentation.

Summary

Cloudflare AI Gateway runs on Cloudflare’s infrastructure in front of the AI providers you use. Its core features, which are analytics, caching and rate limiting, are free on all plans, and you pay for models with your own provider keys or with Cloudflare credits through Unified Billing. Conifer is a hosted gateway that sells usage of more than 230 models from a prepaid balance, serves a named model as named or refuses it, and states the exact cost of every request.

Cloudflare AI Gateway details were read from developers.cloudflare.com and cloudflare.com on 7 November 2026. Plans and prices change; check Cloudflare’s pages before you decide.

Side by side

Topic	Conifer	Cloudflare AI Gateway
Where it runs	A hosted service at one Conifer address; there is nothing to deploy	On Cloudflare’s infrastructure, as part of your Cloudflare account. It is available on all plans, and a default gateway is created on your first request
Model access	More than 230 models paid from one prepaid balance, with no provider accounts needed; your own Azure OpenAI resource can sit behind the same key	Your own keys for providers such as OpenAI, Anthropic, Amazon Bedrock, Azure OpenAI and Google Vertex AI, or Cloudflare credits through Unified Billing. Its model catalog listed 272 models, from Workers AI and third-party providers
Interfaces	OpenAI Chat Completions, OpenAI Responses and Anthropic Messages on one key, plus embeddings	OpenAI Chat Completions, OpenAI Responses and Anthropic Messages endpoints, a universal endpoint for every model type, and each provider’s own API
What you pay	Every model is billed at its maker’s published list price per token, from a prepaid balance with no subscription. A top-up adds a small processing line that covers Stripe’s card fee	Core features are free. Credits bought through Unified Billing carry a 5% fee, and provider token prices pass through with no markup. Logs follow Workers Logs pricing for accounts whose first gateway dates from 24 September 2026 or later, and guardrails are billed as Workers AI usage
Your own provider keys	No Conifer fee on requests served by your own Azure OpenAI resource. AWS Bedrock and Google Vertex are not served as your own endpoints today	Keys for supported providers can be stored in Cloudflare’s Secrets Store or sent with each request. A gateway setting stops requests without your key from falling back to Cloudflare credits
Failover	When a provider fails, rate-limits or times out, the gateway serves the same model from another provider at the same price, for models with more than one provider. You can also list up to three fallback models	Fallbacks are an ordered list of providers or models, tried on an error or a timeout, and a response header shows which step served. Dynamic routes add conditions, budgets and fallbacks that you can change without touching application code
Automatic model choice	The router, in beta, picks one model from a pool of up to 64 models you approve, and does not switch to a different model on its own. Routing adds no fee: a request sent to `beta-auto` costs only the model that answers it	Auto Router picks a model from a pool Cloudflare manages, or from models you choose, and moves to the next-ranked model if a provider cannot serve the request. Response headers name the model it chose
Cost on each request	Every response carries the amount charged, itemised by token type, and a streamed response ends with the same receipt	Logs and analytics show a cost for each request. Cloudflare’s costs page calls the figure an estimate based on token counts, and custom costs can apply negotiated rates
Spending limits	A cost ceiling per request, a spend cap per key and a monthly limit on the account. A request that could exceed a ceiling is refused before any provider is called	Spend limits per gateway, scoped by model, provider or your own metadata such as a user or team, and checked before each request. A burst of concurrent requests can briefly exceed a limit. Rate limits cap requests per time window
Caching	On Anthropic models the gateway caches the stable part of a conversation, so a long system prompt or tool list is paid for once and read at a lower rate after that	Identical requests can be served from Cloudflare’s cache instead of the provider, for up to one month
Request logs	The prompt of a standard request is not stored. A response is held for ten minutes so a retry is not charged twice, and request details such as model, tokens and cost are kept for billing	Logs are on by default and can include the prompt and response. You can turn logging off, or skip storing the request and response bodies for one request. Billing details of Unified Billing requests are always logged
Beyond the gateway	The docs cover the gateway, the router and the SDKs	Guardrails that screen prompts and responses, data loss prevention scanning, log export with Logpush, and custom providers at any HTTPS address
Teams and access	Team accounts with owner, admin and member roles, per-key model scope and spend caps, set up on request	Gateways belong to a Cloudflare account, and calls to its API use Cloudflare API tokens with set permissions. Enterprise customers confirm pricing with their account team

Which one fits

Cloudflare AI Gateway may suit you if: You already run on Cloudflare, you want repeated requests served from a cache, you want guardrails or data loss prevention in the gateway, or you already hold provider keys and want analytics, caching and rate limits in front of them at no charge.
Conifer may suit you if: You want many models paid from one balance at their makers’ list prices, you need each response to state the model that served and the exact amount charged, you want a cost ceiling on each request enforced before any provider is called, or you want automatic model choice that stays within a pool of models you approve.
Using both: Cloudflare AI Gateway accepts custom providers at any HTTPS address. Add Conifer as a custom provider, and Cloudflare’s caching, rate limits and logs sit in front of the models Conifer serves.

Moving from Cloudflare AI Gateway to Conifer

Address and key: Point an OpenAI or Anthropic client at Conifer and use a Conifer key.
OPENAI_BASE_URL=https://api.conifer.build/v1 OPENAI_API_KEY=$CONIFER_API_KEY
Model names: Use the model names in the Conifer catalog. Names with the lab in front, such as anthropic/claude-opus-5, are also accepted.
Fallbacks: Conifer tries another provider for the same model on its own. To try other models, list up to three fallbacks in a request header.
Limits: Give each person or service its own key with a spend cap, and set a cost ceiling on any request. Team accounts are set up by Conifer, and an admin can limit a key to a list of models.
Check first: AWS Bedrock and Google Vertex are not served as your own endpoints today, the moderations endpoint is not served, and image, audio, Files and Batches endpoints have no Conifer equivalent.

Replace your provider key has the full steps, Routing covers fallbacks and failover, and Billing and caching covers spend caps and cost ceilings.

Conifer Gateway: One API key and endpoint for every model in the catalog, with the model and cost stated on every response.
What is an LLM gateway?: How a gateway sits between your software and the companies that serve models, and what to check when you choose one.
Conifer vs OpenRouter: The two gateways side by side, from each company's own documentation.
Conifer vs Portkey: The two gateways side by side, from each company's own documentation.
Conifer vs LiteLLM: The two gateways side by side, from each company's own documentation.
All comparisons: Conifer compared with eleven other LLM gateways and routers, from their own documentation.

Questions

Is Conifer an alternative to Cloudflare AI Gateway?: Yes, for sending requests to models from many providers with fallbacks, spending limits and a cost for each request. Cloudflare AI Gateway also offers response caching, guardrails and data loss prevention, and its core features are free. Conifer sells model usage from a prepaid balance, serves a named model as named or refuses it, and states the exact cost of each request.
Can I use Cloudflare AI Gateway and Conifer together?: Yes. Cloudflare AI Gateway accepts custom providers at any HTTPS address, so you can add Conifer as a provider and keep Cloudflare’s caching, rate limits and logs in front of it.
Do I need provider accounts to use Conifer?: No. One Conifer key covers every model in the catalog, paid from a prepaid balance. You can also put your own Azure OpenAI resource behind the same key, and Conifer adds no fee on that usage.
Does Conifer store my prompts?: The gateway does not store the prompt of a standard request. It holds a response for ten minutes so that a retried request is not charged twice, and keeps request details such as the model, token counts and cost for billing.