Live OpenAI-compatible API · Self-hosted on your server

One Endpoint.
Every AI Model.

Unified API gateway for OpenAI, Anthropic, Claude. Sell AI access with credit billing, Telegram integration, and zero infrastructure headaches.

Get API Key via Telegram View API Docs

Python — OpenAI SDK
from openai import OpenAI
client = OpenAI(
  api_key="biran-your-api-key",
  base_url="
)
response = client.chat.completions.create(
  model="biran/smart",
  messages=[{"role":"user","content":"Hello!"}]
)
# Works with LangChain, LlamaIndex, Vercel AI, any
              SDK

Multi-Provider Routing

Automatically route to the cheapest or best provider. Switch providers without changing your code.

OpenAI

gpt-5.2, gpt-4o,
gpt-4o-mini

Anthropic

claude-opus-4,
claude-sonnet-4

Google

gemini-3.1-pro,
gemini-3-flash

DeepSeek

deepseek-v3.2,
deepseek-r1

+ More

easily extensible

biran/cheapCheapest

biran/smartBalanced

biran/bestTop quality

biran/autoSmart routing

Everything you need to sell AI API

Built for developers who want to monetize AI access. Full billing, no third-party payment required.

Credit-Based Billing

5-hour credit windows with automatic reset via Redis TTL. Zero cron jobs. Fair usage built in automatically.

Smart Model Routing

Aliases like biran/cheap auto-select the best provider. No code changes needed.

OpenAI Compatible

Drop-in replacement for OpenAI API. Works with LangChain, LlamaIndex, Vercel AI, any SDK — zero code changes.

Telegram Integration

User bot: /start, /buy, /apikey. Admin bot: /orders, /confirm. No web dashboard needed to manage users.

Streaming (SSE)

Full streaming support for chat completions. Real-time responses, fully compatible with OpenAI streaming spec.

Rate Limiting

Per-user, per-plan rate limits via Redis. Configurable per tier. Prevents abuse and ensures fair resource sharing.

From zero to API revenue

Three steps to start selling AI API access

User registers via Telegram

Send /start to the bot, choose a plan, make payment. Admin confirms via bot. Done in minutes.

API Key delivered instantly

Bot generates and sends the API key right away. Users can create multiple keys with custom names.

Start building immediately

Use with any OpenAI-compatible SDK. Credits reset every 5 hours. Monitor usage with /usage command.

Simple, transparent pricing

POST /v1/chat/completions

Chat completion

curl -X POST \
-H "Authorization: Bearer biran-YOUR_KEY" \
-H "Content-Type: application/json" \
a-i18n="pricing.sub">Start free, scale as you grow. No hidden fees.

Prices load from GET /public/plans on biran-web (WEB_PORT, default 8080). Set window.BIRAN_WEB_URL to override.

Loading plans…

API Reference

Fully OpenAI-compatible. Just change the base URL.

POST /v1/chat/completions

Chat completion (OpenAI format)

              curl -X POST  \

                -H "Authorization: Bearer biran-YOUR_KEY" \

                -H "Content-Type: application/json" \

                -d '{"model":"biran/smart","messages":[{"role":"user","content":"Hello"}]}'

POST /v1/messages

Chat completion (Anthropic format)

              curl -X POST  \

                -H "x-api-key: biran-YOUR_KEY" \

                -H "anthropic-version: 2023-06-01" \

                -H "Content-Type: application/json" \

                -d '{"model":"claude-sonnet-4","max_tokens":1024,"messages":[{"role":"user","content":"Hello"}]}'

GET /v1/models

List available models

              curl  \

                -H "Authorization: Bearer biran-YOUR_KEY"

Connect with OpenClaw

Use Biran Gateway as your AI backend in OpenClaw — just point the base URL and you're set.

Provider config

Add Biran as a provider in ~/.openclaw/openclaw.json under models.providers.

"biran": {
  "baseUrl": "",
  "apiKey": "biran-YOUR_API_KEY",
  "api": "openai-completions",
  "models": [
    { "id": "biran/smart", "name": "Biran Smart" },
    { "id": "biran/best", "name": "Biran Best" },
    { "id": "biran/cheap", "name": "Biran Cheap" }
  ]
}

Enable in agent defaults

Add model aliases under agents.defaults so OpenClaw can use them.

"agents": {
  "defaults": {
    "models": {
      "biran/biran/smart": { "alias": "smart" },
      "biran/biran/best":  { "alias": "best" },
      "biran/biran/cheap": { "alias": "cheap" }
    }
  }
}

Frequently Asked Questions

Quick answers to common questions about Biran Gateway.

What is Biran Gateway?

Biran Gateway is a self-hosted AI API gateway that provides a unified OpenAI-compatible endpoint for multiple providers (OpenAI, Anthropic, Claude). It includes credit-based billing, Telegram integration, and a web dashboard.

Is it compatible with OpenAI SDK?

Yes. Just change the base_url to your Biran Gateway endpoint. It works with any OpenAI-compatible SDK including LangChain, LlamaIndex, Vercel AI, and the official OpenAI Python/Node libraries.

How does the credit system work?

Each plan has a credit allowance that resets every 5 hours automatically via Redis TTL. No cron jobs needed. Different models cost different amounts of credits per request based on token usage.

How do I get an API key?

Send /start to our Telegram bot, choose a plan, and complete payment. The bot will generate and send your API key instantly after admin confirmation.

What models are available?

We support OpenAI (GPT-5.2, GPT-4o, GPT-4o-mini), Anthropic (Claude Opus 4, Claude Sonnet 4), Google (Gemini 3.1 Pro, Gemini 3 Flash), DeepSeek (V3.2, R1), and more. You can also use smart aliases like biran/cheap, biran/smart, biran/best for automatic routing.

Is there a free plan?

Yes. The free plan gives you a limited number of credits per 5-hour window with basic rate limits. It's great for testing and small projects.

One Endpoint. Every AI Model.