Flex
Lower rates · variable availability
For experiments and flexible workloads. Requests may fail when capacity is limited.
View Flex ratesClaude, GPT and more.
One independent API gateway.
Prepaid in USD. A $20 top-up adds $20 to your balance. Tax is additional and grants no API balance.
Use the tools you already work with. Check the model catalog for availability and rates in your chosen pool.
curl https://api.zrelay.net/v1/messages \
-H "x-api-key: YOUR_ZRELAY_KEY" \
-H "anthropic-version: 2023-06-01" \
-H "content-type: application/json" \
-d '{
"model": "claude-sonnet-5-5",
"max_tokens": 256,
"messages": [
{"role": "user", "content": "Hello"}
]
}'
Use an enabled model from your key’s pool.
Read the API guide
Requests pass through Zrelay to inference suppliers. Read about data processing.
Flex keeps costs down. Prime offers more consistent availability at higher rates. Choose a pool per key; all use your shared balance. Xtra adds focused model catalogs. Check current availability before sending a request.
Lower rates · variable availability
For experiments and flexible workloads. Requests may fail when capacity is limited.
View Flex ratesHigher rates · more consistency
For workloads where interruptions disrupt your work. Review current model availability.
View Prime ratesFocused fast models
GLM and DeepSeek variants. Check current model availability and rates.
View Xtra Flash ratesFocused abliterated Kimi model
kimi-k31-pro-abliterated. Check current model availability and confirmed catalog rates.
View Xtra Prime Kimi ratesModel availability depends on your pool. Requests stay there by default. You can enable Flex → Prime fallback per key for verified equivalent models, at Prime rates using general balance. How Zrelay works
Understand your balance, pools and payments before making your first paid request.
How Zrelay worksNo. Your paid keys share one USD balance across model families and pools. Each request is charged at the rates of its selected pool.
Yes. Select an eligible pool and model family for an API key in your account and save the change. You can separately enable Flex → Prime fallback, off by default. A verified equivalent may receive one Prime attempt before response content starts, using only general balance at Prime rates.
Choose an amount in USD in your account and an available payment method. A $20 purchase of API balance adds $20 of usage balance; tax, if applicable, is added separately and grants no usage balance. Check the final total before paying. Payment availability is shown before checkout. See the refund policy for payment support and additional information in Spanish.
Find the right endpoint, configure your client, or get help with your account.