API Documentation

API overview #

OpenAI-compatible gateway for chat, embeddings, and agent harness integrations. All client requests use the same base URL and sk-ixdm- API key.

{
  "status": "ok",
  "database": "up"
}

Quickstart #

Get started in three steps:

from openai import OpenAI

client = OpenAI(
    base_url="https://ai.ixdm.co.id/v1",
    api_key="sk-ixdm-xxxx"
)

response = client.chat.completions.create(
    model="deepseek-chat",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://ai.ixdm.co.id/v1',
  apiKey: 'sk-ixdm-xxxx'
});

const response = await client.chat.completions.create({
  model: 'deepseek-chat',
  messages: [{ role: 'user', content: 'Hello!' }]
});
console.log(response.choices[0].message.content);

curl https://ai.ixdm.co.id/v1/chat/completions \
  -H "Authorization: Bearer sk-ixdm-xxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-chat",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Authentication & keys #

Every request must include a valid API key in the Authorization header:

Authorization: Bearer sk-ixdm-xxxx

API keys are generated from the user dashboard or assigned by an admin. Keys begin with the sk-ixdm- prefix.

Chat completions #

Send a chat completion request. The router selects the appropriate provider and upstream key based on the model and your client configuration.

Request parameters

Response

{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "model": "deepseek-chat",
  "choices": [{
    "index": 0,
    "message": {"role": "assistant", "content": "Hello! How can I help?"},
    "finish_reason": "stop"
  }],
  "usage": {"prompt_tokens": 10, "completion_tokens": 8, "total_tokens": 18}
}

Streaming #

Set "stream": true in your request to receive Server-Sent Events (SSE). Each chunk is delivered as a data: line.

from openai import OpenAI

client = OpenAI(
    base_url="https://ai.ixdm.co.id/v1",
    api_key="sk-ixdm-xxxx"
)

stream = client.chat.completions.create(
    model="deepseek-chat",
    messages=[{"role": "user", "content": "Hello!"}],
    stream=True
)
for chunk in stream:
    delta = chunk.choices[0].delta
    if delta.content:
        print(delta.content, end="", flush=True)

curl -N https://ai.ixdm.co.id/v1/chat/completions \
  -H "Authorization: Bearer sk-ixdm-xxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-chat",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": true
  }'

Embeddings #

Generate vector embeddings for text input. Uses the same authentication and model routing as chat completions.

Request parameters

curl https://ai.ixdm.co.id/v1/embeddings \
  -H "Authorization: Bearer sk-ixdm-xxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": "The food was delicious."
  }'

Tool calling #

Agent harnesses (Cursor, OpenCode, Hermes) rely on OpenAI-style function calling. Send tools in the request; the model may respond with tool_calls that your client executes and reports back via role: "tool" messages.

Request fields

When the model invokes a tool, finish_reason is "tool_calls":

{
  "choices": [{
    "message": {
      "role": "assistant",
      "content": null,
      "tool_calls": [{
        "id": "call_abc123",
        "type": "function",
        "function": {
          "name": "read_file",
          "arguments": "{\"path\": \"main.go\"}"
        }
      }]
    },
    "finish_reason": "tool_calls"
  }]
}

Models #

Use model aliases configured by the operator. What you see in GET /v1/models depends on your key type:

Common alias examples:

Returns only models your client is authorized to use.

Reasoning #

Some models support extended thinking via the reasoning_effort parameter:

When reasoning_effort is set, the model allocates a portion of its output budget to internal reasoning before producing the final response.

Agent integrations #

Connect AI coding agents and editors using your sk-ixdm- API key and the OpenAI-compatible base URL. All integrations below use the same credentials.

Anthropic SDK harnesses (future) #

Credits & billing #

User accounts use a credit balance denominated in USD to access models. Each model is sold as a credit alias — you must be assigned an alias before you can call its slug.

When your balance is too low for a request, the API returns 402 Payment Required:

{
  "type": "insufficient_credits",
  "message": "insufficient credits. Please top up your balance and try again.",
  "code": "insufficient_credits"
}

Rate limits #

Per-key request rate limits are applied in addition to credit billing. For credit balance and model access, see Credits & billing above.

Errors #

Errors return a JSON envelope:

{
  "type": "invalid_request_error",
  "message": "Missing required parameter: model",
  "request_id": "req-abc123"
}

Status codes