One key for GPT, Claude, Gemini, DeepSeek, Qwen and more — through the OpenAI and Anthropic SDKs your agents already speak, on plans worth up to 30× their price.
import OpenAI from "openai"; const client = new OpenAI({ baseURL: "https://api.xkiro.com/v1", apiKey: process.env.XKIRO_API_KEY,}); const res = await client.chat.completions.create({ model: "anthropic/claude-opus-5.5", messages: [{ role: "user", content: "Hello!" }],}); console.log(res.choices[0].message.content);$ curl https://api.xkiro.com/v1/chat/completions \
-H "Authorization: Bearer $XKIRO_API_KEY" \
-d '{"model":"anthropic/claude-opus-5.5",…}'
✓ 200 OK
"model": "anthropic/claude-opus-5.5"Counted live from the model catalog · updated hourly
Models from
OpenAI
Anthropic
Google
DeepSeek
Qwen
Grok
Kimi
Mistral
MiniMax
MiMo
NVIDIA
Tencent
Cohere
SenseNovaBenefits
One account replaces a stack of provider accounts, invoices and retry code.
the usage for your money on a plan — Max is $20 a month for up to $600 at list price. See the math
free models on every account, no credit card — 500K tokens a day, 1M after Telegram verification.
models from 14 providers behind one key. Switch models by changing one string.
lines of retry code: when a model has more than one route, a failing upstream is retried on the next route automatically.
for requests that fail with a 5xx error — and retrying an identical blocking request replays the first result instead of charging twice.
agent tools with setup guides — Claude Code, Codex, Cline, Cursor, and OpenCode — plus anything that takes a custom base URL.
Plan value
Every paid plan includes up to 30× its price in usage, metered at list price. Max is $20 a month — about $0.67 a day.
You pay$20/ month
≈ $0.67 a day
You can use, up to$600/ month
Metered at catalog list prices
The same 30× on every plan, from $5 to $200 a month. Compare plans
| Model | On a plan, from |
|---|---|
Opus 5.5Up to −96% | $0.14 / $0.67 |
GPT-6.1 SolUp to −96% | $0.067 / $0.34 |
Gemini 3.8 FlashUp to −96% | $0.025 / $0.13 |
Grok 4.7Up to −96% | $0.067 / $0.2 |
MiniMax M3.1 FlashUp to −96% | $0.005 / $0.02 |
Per 1M tokens, input / output. Plan prices assume you use your whole allowance; it's split into 5-hour and 7-day windows (Max: $20 per 5 hours, $140 per 7 days) and unused allowance doesn't roll over.
Pay-as-you-go prices match the official API prices of Anthropic, OpenAI, Google, xAI, and MiniMax (checked Oct 21, 2026).
Slide to what you spend on AI APIs each month at list prices.
Estimate at list prices, assuming your usage fits the plan's 5-hour and 7-day windows. Wallet-billed models aren't covered by plans.
Smallest plan that covers it
Max — $20/month
Covers up to $600 of usage a month
Save up to $580 a month
96% less than paying $600 at list price
Quickstart
No new SDK and no rewrites. Point the client you already use at xKiro — every model in the catalog is one model ID away.
Sign up and create an API key in the dashboard. Free models work right away — no card required.
OpenAI SDKs use api.xkiro.com/v1. Anthropic SDKs and Claude Code use api.xkiro.com — the SDK adds /v1 itself.
Model IDs carry the vendor prefix. Swap one string to move between GPT, Claude, Gemini and the rest.
import OpenAI from "openai"; const client = new OpenAI({ baseURL: "https://api.openai.com/v1", baseURL: "https://api.xkiro.com/v1", apiKey: process.env.XKIRO_API_KEY,}); const res = await client.chat.completions.create({ model: "openai/gpt-6.1-sol", messages: [{ role: "user", content: "Hello!" }],});Your tools
Claude Code, Codex, Cline — and anything that accepts a custom base URL and API key. The config below is copied from our setup guides.
Also set up in minutes:
Catalog
Live per-million-token prices from the catalog. On a plan, usage comes out of your allowance, so the same tokens cost from about 1/30 of the list price.
| Model | Context | Input / 1M | Output / 1M |
|---|---|---|---|
| 1.05M | $0.34 | $1.70 |
| 1M | $0.14 | $0.67 |
| 1.05M | $0.2828 | $0.5655 |
| 1M | $0.025 | $0.13 |
| 500K | $0.067 | $0.2 |
| 262K | $0.1 | $0.5 |
| 1M | Free | Free |
| 1M | $0.005 | $0.02 |
Counted live from the model catalog · updated hourly
Browse all 103 modelsUnder the hood
Follow one request through xKiro. Failover, streaming, fair billing, budgets and logs are part of the path — not add-ons.
When a model has more than one route, a failing upstream is retried on the next route automatically — no retry code on your side.
Answers arrive as they are generated, in the OpenAI or Anthropic streaming format your client already parses.
Metered at catalog rates. The rules are simple and written down:
Plans spend from a 5-hour and a 7-day allowance at catalog rates — watch both fill up in real time.
Model, tokens, cost and status for every call in your dashboard — and as many API keys as you need, each revocable on its own.
Images, speech and web search run on the same key and the same account.
Pricing
Every paid plan includes usage worth up to 30× its price at catalog rates. Prefer no subscription? Pay as you go from your wallet.
The starter plan for individuals getting started with AI routing.
Usage worth up to
$150/mo30× what you pay
Billed monthly
Cancel anytime
Usage budget
$5 every 5 hours · $35 every 7 days
For daily development — 4× the usage of Pro.
Usage worth up to
$600/mo30× what you pay
Billed monthly
Cancel anytime
Usage budget
$20 every 5 hours · $140 every 7 days
For power users — 20× the usage of Pro.
Usage worth up to
$3,000/mo30× what you pay
Billed monthly
Cancel anytime
Usage budget
$100 every 5 hours · $700 every 7 days
More plans
For regular use — 2× the usage of Pro.
Billed monthly
Usage worth up to
$300/mo30× what you pay
For teams and large workloads — 40× the usage of Pro.
Billed monthly
Usage worth up to
$6,000/mo30× what you pay
Free·$0 forever—45 free models · 500K tokens a day · no card
Start free| All prices in USD | Free | Pro | Pro+ | RecommendedMax | Ultra | Power |
|---|---|---|---|---|---|---|
$0 | $5/mo | $10/mo | $20/mo | $100/mo | $200/mo | |
| Free forever | Billed monthly | Billed monthly | Billed monthly | Billed monthly | Billed monthly | |
| Start | Choose | Choose | Choose | Choose | Choose | |
| Usage budget | ||||||
| Every 5 hours | — | $5 | $10 | $20 | $100 | $200 |
| Every 7 days | — | $35 | $70 | $140 | $700 | $1,400 |
| Worth up to / month | — | $15030× what you pay | $30030× what you pay | $60030× what you pay | $3,00030× what you pay | $6,00030× what you pay |
| Free-model tokens / day | 500K1M after Telegram verification | 8M | 16M | 32M | 160M | 320M |
| Models | ||||||
| Free models | All 45 | All 45 | All 45 | All 45 | All 45 | All 45 |
| Paid models in your budget | — | 53 | 53 | 53 | 55 | 55 |
| Smart routing | — | Included | Included | Included | Included | Included |
| Automatic fallback on errors | — | Included | Included | Included | Included | Included |
| Capacity & support | ||||||
| AI image generation (GPT Image) | — | Basic | More | High | Very high | Max |
| Concurrent requests | Low | Standard | High | Higher | Very high | Max |
| Priority at peak traffic | — | — | — | — | Included | Included |
| Support | Community | Email · 24h | Priority email | Priority | High priority | Highest priority |
Usage budget
Models
Capacity & support
No plan needed. Top up your wallet and pay per request at catalog rates — free models stay free.
Top up in the dashboard →Higher limits for a team or a production workload — tell us what you need.
Contact us →Free plan
No card and no trial clock — the free models stay free every day. Move to a plan when you want usage worth up to 30× what you pay.
No credit card · 45 free models, free forever
Live event
Verify Telegram while the event runs to get one Welcome Drop box. It always pays out between 50M and 100M tokens.
How to get it
Ends in
Ends Nov 22, 2026, 00:00 GMT+7
Tokens work on MiniMax models and expire 14 days after you receive them. One per email and Telegram account, after an anti-fraud check.
Already a member? Open the event →FAQ
How plans, free models and billing actually work.
Still have a question? Contact us →Yes. 45 models are free to call on every account, within a daily allowance of 500K tokens — 1M after you verify on Telegram. No credit card required.
The OpenAI and Anthropic SDKs, and anything built on them. Use https://api.xkiro.com/v1 as the base URL for OpenAI clients and https://api.xkiro.com for Anthropic clients such as Claude Code, then pass any model ID from the catalog.
Each paid plan includes a spending allowance over a 5-hour and a 7-day window, measured at catalog rates. Usage inside both windows is covered by the plan; unused budget does not carry over to the next window.
A plan is a fixed monthly price with usage budgets. Pay-as-you-go draws from a prepaid wallet balance, charged per request at catalog rates — used when you have no plan, and for the few models billed outside plan budgets.
No. Requests that fail with a 5xx error or never reach a provider are not charged. If you cancel a stream, you pay only for the tokens generated before you stopped.
When a model has more than one route, xKiro retries on the next route automatically, so a single upstream outage does not have to fail your request.
Yes. Cancel auto-renewal at any time from Billing and keep access until the end of the period. You can upgrade from the pricing page in your dashboard.
Yes. All plans, including the free tier, allow commercial use. Usage is also subject to each model provider's terms.
Create an API key and start with 45 free models — move to a plan when you want usage worth up to 30× what you pay.
No credit card · Cancel anytime · 45 free models, free forever
Cursor
OpenCode
