# TokenFajri — Enterprise AI API Gateway Indonesia

> TokenFajri (https://token.fajriag.ai.id) is an enterprise-grade, OpenAI-compatible AI API Gateway in Indonesia providing unified access to Anthropic Claude, OpenAI GPT, DeepSeek, Google Gemini, Meta Llama, and Mistral models through a single endpoint (`https://token.fajriag.ai.id/v1`) with instant domestic IDR (Rupiah / QRIS) prepaid billing and zero prompt retention.

## Core Machine-Readable Endpoints for AI Agents
- Full LLM Documentation: https://token.fajriag.ai.id/llms-full.txt
- OpenAPI 3.1 Specification: https://token.fajriag.ai.id/openapi.json
- AI Plugin Manifest: https://token.fajriag.ai.id/.well-known/ai-plugin.json
- OpenAI v1 Base URL: https://token.fajriag.ai.id/v1
- Chat Completions Endpoint: POST https://token.fajriag.ai.id/v1/chat/completions
- Model Registry Endpoint: GET https://token.fajriag.ai.id/v1/models

## Key Capabilities & Value Proposition
- **100% OpenAI SDK Compatible**: Drop-in replacement for OpenAI Python SDK, OpenAI TypeScript/Node SDK, Cursor IDE, VS Code Cline, Roo Code, LangChain, and Vercel AI SDK. Simply set `baseURL` to `https://token.fajriag.ai.id/v1` and pass `Authorization: Bearer sk-tok-...`.
- **Domestic Indonesian Rupiah (IDR) Settlement**: Instant prepaid top-up and subscription licensing via QRIS and Indonesian bank transfer with 0% USD foreign exchange fees.
- **Multi-Pool Auto-Failover**: Automatic upstream credential rotation on HTTP 429/503 with <2ms gateway routing overhead.
- **Zero Prompt Retention**: Direct Server-Sent Events (SSE) pass-through streaming; prompt and completion payloads are never stored on disk.

## Live Active AI Models (`https://token.fajriag.ai.id/v1/models`)
- `deepseek-v4.1-flash` (DeepSeek — DeepSeek V4.1 Flash): FREE TIER (Rp 0) | Context: 1,000,000 tokens
- `mimo-v2.6-flash:free` (th — mimo v2.6 Flash:free): FREE TIER (Rp 0) | Context: 1,048,576 tokens
- `muse-spark-1.2-contributor-free` (OpenCode — muse spark 1.2 contributor free): FREE TIER (Rp 0) | Context: 128,000 tokens

## Licensing & Request Quota Packages (IDR)
- **Paket Sprint 3 Hari** (WEEKLY): Rp 10.000 for 200 requests (3 days) — 200 Request AI aktif selama 3 hari. Pas untuk uji coba atau sprint proyek kilat.
- **Paket Mingguan 7 Hari** (WEEKLY): Rp 25.000 for 500 requests (7 days) — 500 Request AI aktif selama 7 hari. Pas untuk kebutuhan coding mingguan.
- **Paket Bulanan 1 Bulan** (MONTHLY): Rp 85.000 for 3,000 requests (30 days) — 3.000 Request AI aktif selama 30 hari (1 Bulan). Paling hemat untuk developer harian.
- **Paket Triwulan 3 Bulan** (MONTHLY): Rp 240.000 for 10,000 requests (90 days) — 10.000 Request AI aktif selama 90 hari (3 Bulan). Solusi hemat untuk jangka menengah.
- **Paket Semester 6 Bulan** (MONTHLY): Rp 450.000 for 22,000 requests (180 days) — 22.000 Request AI aktif selama 180 hari (6 Bulan). Nilai terbaik untuk developer & tim profesional.

## Quick Start for AI Coding Agents (Cursor / Cline / SDK)
```typescript
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://token.fajriag.ai.id/v1",
  apiKey: process.env.TOKEN_API_KEY, // Obtain at https://token.fajriag.ai.id/register
});

const response = await client.chat.completions.create({
  model: "claude-sonnet-4-5",
  messages: [{ role: "user", content: "Hello from AI Agent" }],
  stream: true,
});
```

## Navigation & Support
- Website & Live Sandbox: https://token.fajriag.ai.id/
- Developer Registration: https://token.fajriag.ai.id/register
- Console Login: https://token.fajriag.ai.id/login
- Support Telegram: https://t.me/FAJRIAG1