Pay 40× less
for vibe coding
Connect Codex, Cursor, Crush or any SDK in two clicks — get access to every top model through a single key.
GPT
Claude
Grok
Qwen
Kimi
DeepSeek
GLM
Gemini
MiMo
MiniMax
One key. Zero routine.
One key for every model
Build your workflow anywhere, any way you like. Top up once, plug in, ship.
Unified billing
Spending across models, keys and projects in one dashboard. You are billed only for actual tokens.
Universal API
Works with OpenAI SDK, Codex, Cursor, VS Code and any client that supports a base_url.
Flagship, fast and cheap
GPT, Claude, Grok, Qwen, Kimi, DeepSeek, GLM, Gemini, MiMo, MiniMax and more.
Pay any way you like
No foreign cards, no VPN, no subscriptions you always forget to cancel.
Support on the line
If anything breaks while connecting, our support team is always one message away.
Cheaper than direct API access
The margin of direct access is the inconvenience: separate accounts, currencies, manual routing. We merge it all and show you the price.
Works with every client
Your tools. Your workflow. One API key.
CrushTerminal agent
Claude CodeCoding agent
CodexCoding agent
CursorAI code editor
VS CodeEditor extensions- OpenAI SDKBuild your own
- Anthropic SDKBuild your own
- Your clientJust set
base_url
Questions? Let's solve them
Are these the same models?
Yes. You send requests to a compatible gateway and pick any model from the list in the model field.
Does it work with Codex, Cursor and the OpenAI SDK?
Yes — just point the gateway address in base_url and use your key. Setup takes under a minute.
Do you store prompts?
We do not collect prompt or response contents. Logs keep only technical data for billing and diagnostics.
Start saving today.
Registration takes under a minute. The first key plugs straight into Codex, Cursor or any SDK.
Pick a modelThe value table
All 37 models behind one key, with honest per-model pricing and the tokens you get per dollar. Sorted by value — pick your weapon.
| Model | $/1M tokens | Tokens per $1 | Context | Max out | Skills |
|---|
Effective rates at the best top-up tier ($200+) with the current markup. Multipliers shown as billed by the gateway.
Welcome back
Sign in with your passkey — no passwords here.
Hi, user
My API key
Rotation invalidates the previous key instantly.
How tokens are charged
billed = (input + output) × model multiplier. Example: (61 + 192 + 39) × 0.05 = 14 for mimo-v2.5. Cache reads are counted as input.
Full request logs
| Date | Model | Actual | Billed |
|---|
Top up with crypto
Send USDT (TRC-20 / BEP-20 / ERC-20) or native TRX / BNB / ETH to your personal address. Detected transfers land here after network confirmation; an admin credits tokens at the current rate.
Converter — crypto to tokens
loading live prices…
Your requests
| Created | Network | Asset | Amount | Locked | Status |
|---|
Deposit history
| Date | Network | Asset | Amount | Status | Tokens |
|---|
Support tickets
Connection guide
One base URL, one key — every client below works out of the box. Replace https://your-domain with this site address and sk-px-… with your key.
Which API styles are supported?
OpenAI-compatible /v1/chat/completions and /v1/responses, Anthropic-compatible /v1/messages plus /v1/messages/count_tokens, and GET /v1/models / GET /v1/balance.
How is billing calculated?
billed = (input + output tokens) × multiplier of the model. Multipliers are listed on the Models page.
Is streaming supported?
Yes — SSE streaming works on chat completions and messages endpoints.
Account settings
Passkeys
Add a passkey as a secondary sign-in method (fingerprint / face / hardware key).
Two-factor (TOTP)
status: unknown
Lost your hash code?
The code is shown once at registration and stored only as a hash. If you lost the file — open a support ticket from a signed-in session and we will verify you manually.
Admin console
Top users by spend
| User | Tokens |
|---|