TokenZero vs Cloudflare AI Gateway
Cloudflare AI Gateway is an edge proxy with caching, rate limiting, logs and analytics in front of model providers. The matrix below says what each product is and how it bills; the only numbers are TokenZero's own measurements and prices.
TokenZero and Cloudflare AI Gateway differ in kind: Cloudflare AI Gateway is edge AI gateway; TokenZero is a BYOK gateway whose engines reduce the provider bill and sign a receipt per call. Measured overhead is 1.52 ms p50; the cascade holds a 99.8% parity floor.
· Source: Cloudflare AI Gateway documentation at https://developers.cloudflare.com/ai-gateway/ read 2026-10-26; TokenZero overhead from /benchmarks; prices from table 2026-10-26
At a glance
- Compared with
- Cloudflare AI Gateway (edge AI gateway)
- TokenZero overhead
- 1.52 ms p50 / 3.42 ms p95 (measured, /benchmarks)
- BYOK
- TokenZero: always · Cloudflare AI Gateway: yes
- Receipts
- TokenZero: signed per call · Cloudflare AI Gateway: none
- Billing
- TokenZero: 15% of verified savings · Cloudflare AI Gateway: included with Cloudflare plans; log volume by plan
- Reference saving (gpt-6.1-sol, 100k calls)
- $285/month from prefix caching (computed)
- Net after 15% fee
- $242/month (computed)
- Parity floor
- 99.8%
- Free tier
- $500/month routed
- Vendor page read
- 2026-10-26
TokenZero vs Cloudflare AI Gateway: what each one is (vendor page read 2026-10-26)
| Dimension | TokenZero | Cloudflare AI Gateway |
|---|---|---|
| Category | BYOK savings gateway | edge AI gateway |
| Hosting | hosted | hosted (Cloudflare) |
| Bring your own key | always; keys pass through and are never stored | yes |
| Open source | CLI, MCP server, audit core and GitHub Action Apache-2.0; gateway hosted | closed |
| How it bills | free up to $500/month routed; then 15% of verified savings, 3x-ROI-or-$0, capped plans | included with Cloudflare plans; log volume by plan |
| Caching | prompt-cache formatting + exact and semantic response cache | response cache at the edge |
| Routing | cascade to the cheaper sibling inside one provider, 99.8% parity floor, advisory by default | fallback and universal endpoint you configure |
| Savings receipts | Ed25519-signed per call, hash-chained, public key published | none |
| Gateway overhead | 1.52 ms p50 / 3.42 ms p95 (measured, /benchmarks) | not measured by TokenZero |
| Data retention | metadata only; ZDR mode available | see vendor docs |
| Net saving, 100k gpt-6.1-sol calls/month (prefix caching only) | $242 after fee | — |
What you keep when you move from Cloudflare AI Gateway
the base-URL swap; TokenZero sits behind Cloudflare too.
- Response caching is one of four engines: prompt-cache formatting and cascade routing cut requests that cannot be cached at all.
- Cloudflare's User Insights (changelog 2026-08-05) adds per-user spend visibility and anomaly detection in the dashboard; TokenZero acts in-flight on each request, injecting prefix-cache breakpoints and cascading simple prompts under the 99.8% parity floor.
- Savings are signed per request against the list-price baseline, not inferred from hit rates.
- Hosted audits and the CLI find waste in source before it is sent.
The change
OPENAI_BASE_URL=https://tokenzero.sh/v1
ANTHROPIC_BASE_URL=https://tokenzero.sh
# keep your provider keys exactly as they areThen npx -y @tokenzero/cli doctor to confirm the gateway, signer and pricing table, and see the switch guide.
Questions people ask
What is the difference between TokenZero and Cloudflare AI Gateway?
Cloudflare AI Gateway is an edge proxy with caching, rate limiting, logs and analytics in front of model providers. TokenZero is a BYOK gateway: you keep the base-URL swap; TokenZero sits behind Cloudflare too, and every call comes back with a signed receipt. Response caching is one of four engines: prompt-cache formatting and cascade routing cut requests that cannot be cached at all.
What does switching from Cloudflare AI Gateway to TokenZero save?
It depends on your prompts; on the reference workload (100,000 calls of 2,000 input with a 1,500-token static prefix and 300 output on gpt-6.1-sol, table 2026-10-26) prefix caching alone cuts $700 to $415 a month, $285 saved and $242 net after the 15% fee; exact-cache hits and cascade routing add to that. The receipts show the real figure.
Can I keep Cloudflare AI Gateway and add TokenZero?
Usually yes: chain them by pointing Cloudflare AI Gateway's upstream at https://tokenzero.sh (or TokenZero's X-TokenZero-Upstream at Cloudflare AI Gateway); the receipt still prices the request at list and records what the gateway did.
How is the 15% fee computed?
On verified savings only: list-price baseline minus provider-reported actual cost, per signed receipt. A cycle whose savings are below 3x the fee is billed $0; Pro Cap and Scale Cap add monthly ceilings. The first $500 of routed spend each month is free.