Every frontier model.
One endpoint.
APIBrain is an OpenAI-compatible gateway. One key, one base URL, thirteen live models — point the tools you already use at it and switch models with a single string.
| Time | Status | Model | Tokens | User |
|---|---|---|---|---|
| Loading live activity… | ||||
Live requests through the gateway — masked, aggregate. No names, keys, or payloads.
One base URL
Every model answers on the same endpoint. Switching model is one string, not one migration.
Drop-in client
Anything that already speaks the OpenAI chat API works untouched — VS Code, Cursor, your own scripts. Streaming included.
Token allowance
Your pack sets the ceiling. The dashboard shows exactly what's left, and calls stop cleanly at zero.
Live in three steps
No sales call and no onboarding queue. Redeem, copy, call.
Redeem your code
Apply the code on your dashboard. It sets the allowance attached to your account.
Copy your key
Create a key and copy it once. Keys can be revealed again or revoked at any time.
Change one line
Point your client at api.apibrain.top/v1. Everything else stays as it is.
13 models, one key
Reasoning, coding, long-context and vision — all behind the same address, all on the same key.
Token packs, not subscriptions
Every pack is a block of tokens valid for 30 days. Stack them any time — nothing auto-renews.
Starter
- 6,000,000 tokens
- 30-day validity
- Every live model
Developer
- 20,000,000 tokens
- 30-day validity
- Every live model
- Usage dashboard
Heavy coding
- 50,000,000 tokens
- 30-day validity
- Every live model
- Usage dashboard
Max
- 100,000,000 tokens
- 30-day validity
- Every live model
- Lowest rate per token
OpenAI-compatible, so nothing else changes
If your tool already speaks the OpenAI API, you change two values and keep every line of your code.
from openai import OpenAI client = OpenAI( base_url="https://api.apibrain.top/v1", api_key="GW-••••••••", ) r = client.chat.completions.create( model="deepseek-v4.1-flash", messages=[{"role": "user", "content": "Hello"}], ) print(r.choices[0].message.content)
import OpenAI from "openai"; const client = new OpenAI({ baseURL: "https://api.apibrain.top/v1", apiKey: "GW-••••••••", }); const r = await client.chat.completions.create({ model: "deepseek-v4.1-flash", messages: [{ role: "user", content: "Hello" }], }); console.log(r.choices[0].message.content);
curl https://api.apibrain.top/v1/chat/completions \ -H "Authorization: Bearer GW-••••••••" \ -H "Content-Type: application/json" \ -d '{ "model": "deepseek-v4.1-flash", "messages": [{"role":"user","content":"Hello"}] }'
# point any OpenAI-compatible tool at APIBrain OPENAI_BASE_URL=https://api.apibrain.top/v1 OPENAI_API_KEY=GW-•••••••• # same tools, same code, different door
What you get
Everything below is served by the gateway. You never talk to a model provider directly.
- Streaming —
stream: trueworks as-is - Vision input on supported models
- Long context — up to 1M tokens
- Usage per key — tokens, requests, spend
- Revocable keys — rotate any time
Questions, answered
Is it really OpenAI-compatible?
Yes. Same request shape, same response shape, same streaming format. Any client, library or tool built for the OpenAI API works by changing the base URL and the key — nothing else.
How do I get access?
Access is by invitation while we scale. Request an invite above; when a spot opens we'll email you a one-time invitation code. Redeem it on the dashboard to create your account — the same code sets the token allowance attached to your first key.
Can I use it from VS Code, Cursor or my own scripts?
Yes — those are the main use cases. Anything that accepts a custom OpenAI base URL works.
What happens when my tokens run out?
Calls are refused with a clear error until you redeem another code. Nothing renews automatically and you're never charged past your allowance.
Do you keep my prompts?
Usage is recorded as counts and tokens so your allowance and dashboard work. Your prompts and responses are not published or shared, and nothing about your account appears on this page.
Where do I see my usage and live status?
Behind login, on api.apibrain.top. The stat strip above is public aggregates refreshed every five minutes; the live activity table below refreshes every few seconds.
One key. Every model.
Access is by invitation while we scale. Existing accounts can sign in; new accounts need an invite from the waitlist.
Sign in