Unified API for every
AI model

Route requests to Claude, GPT, Gemini, DeepSeek through one endpoint. OpenAI and Anthropic SDK compatible. Pay per token.

Powered by
Claude
GPT
Gemini
DeepSeek
$ curl https://ai.osiris-code.com/v1/chat/completions \
  -H "Authorization: Bearer sk-osiris-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4.5",
    "messages": [{"role": "user", "content": "Hello"}]
  }'
32
Models available
2
Upstream providers
3
API formats
<200ms
Routing overhead

Supported models

Claude Opus 4.6 Anthropic
Claude Sonnet 4.5 Anthropic
GPT-5.4 OpenAI
GPT-5.2 Codex OpenAI
Gemini 3.1 Pro Google
Gemini 2.5 Flash Google
DeepSeek V3.2 DeepSeek
Qwen3 Coder Alibaba

View all 32 models including pricing multipliers.

How it works

1

Create an account

Sign up and get your API key. No credit card required.

2

Top up tokens

Buy a token package. Tokens never expire. Use them whenever.

3

Send requests

Point your OpenAI or Anthropic SDK to our endpoint. That's it.

Drop-in compatible

Change one line in your existing code. Works with OpenAI and Anthropic SDKs.

curl https://ai.osiris-code.com/v1/chat/completions \
  -H "Authorization: Bearer sk-osiris-..." \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4.5",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Pricing

Pay per token. No subscriptions. Tokens never expire.

Starter
100K tokens
Rp 10.000
  • All 32 models
  • Never expires
Get started
Basic
500K tokens
Rp 45.000
  • All 32 models
  • Never expires
Get started
Popular
Pro
1M tokens
Rp 80.000
  • All 32 models
  • Never expires
Get started
Enterprise
5M tokens
Rp 350.000
  • All 32 models
  • Never expires
Get started

Token cost varies by model. Claude Opus uses 3.3x, Haiku uses 0.3x. See multipliers

Multi-provider failover

Requests automatically route to the best available provider. If one is down, we failover instantly.

OpenAI + Anthropic format

/v1/chat/completions, /v1/messages, and /v1/responses. Use whichever format your app already uses.

Pre-deduct billing

Tokens are estimated and deducted before the request. Unused tokens are refunded automatically.