Chat completions

POST /api/v1/llm/v1/chat/completions

OpenAI-compatible chat completions. Supports Idempotency-Key. Every field is in the reference.

Scope: llm:access or llm:chat

Request

Field	Type	Notes
model	string	nebula-4.5 (default) or nebula-v for images.
messages	array, required	system, user, assistant and tool messages. Content is text or content parts.
max_tokens	integer	Output ceiling. An explicit value is clamped to the platform maximum (32,000 by default). max_completion_tokens is an alias; the smaller wins.
temperature	number	0 to 2. Default 0.7.
top_p	number	0 to 1. Default 1.
stream	boolean	Server-sent events. See Streaming.
stop	string or array	Stop sequences.
tools	array	OpenAI function tool definitions.
tool_choice	string or object	auto, none, required, or a named function.
parallel_tool_calls	boolean	Set false for at most one tool call per turn.
response_format	object	{"type": "json_object"} or a json_schema.
reasoning_effort	string	none, low, medium, high, max, xhigh or ultra. Omit for adaptive routing.
user	string	A stable session or end-user id, sent on every turn, so prompt caching can reuse context.

Basic request

cURL

cURLCopy

curl https://api.breachline.io/api/v1/llm/v1/chat/completions \
  -H "Authorization: Bearer $BREACHLINE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nebula-4.5",
    "messages": [
      {"role": "system", "content": "You triage web findings."},
      {"role": "user", "content": "Is a missing HttpOnly flag on a session cookie high severity?"}
    ]
  }'

Python

PythonCopy

response = client.chat.completions.create(
    model="nebula-4.5",
    messages=[
        {"role": "system", "content": "You triage web findings."},
        {"role": "user", "content": "Is a missing HttpOnly flag on a session cookie high severity?"},
    ],
)
print(response.choices[0].message.content)

TypeScript

TypeScriptCopy

const response = await client.chat.completions.create({
  model: 'nebula-4.5',
  messages: [
    { role: 'system', content: 'You triage web findings.' },
    { role: 'user', content: 'Is a missing HttpOnly flag on a session cookie high severity?' },
  ],
});
console.log(response.choices[0].message.content);

Tool calling

Describe your functions in tools. When the model wants one, the response has finish_reason: "tool_calls" and a tool_calls array. Run the tool yourself, then send the result back as a tool message with the matching tool_call_id.

Tool callingCopy

tools = [{
    "type": "function",
    "function": {
        "name": "lookup_asset",
        "description": "Return the owner and environment of a hostname.",
        "parameters": {
            "type": "object",
            "properties": {"hostname": {"type": "string"}},
            "required": ["hostname"],
        },
    },
}]

first = client.chat.completions.create(
    model="nebula-4.5",
    messages=[{"role": "user", "content": "Who owns staging-api.example.com?"}],
    tools=tools,
)
call = first.choices[0].message.tool_calls[0]
result = {"owner": "payments team", "environment": "staging"}

final = client.chat.completions.create(
    model="nebula-4.5",
    messages=[
        {"role": "user", "content": "Who owns staging-api.example.com?"},
        first.choices[0].message,
        {"role": "tool", "tool_call_id": call.id, "content": json.dumps(result)},
    ],
    tools=tools,
)

Structured output

Ask for JSON that matches a schema with response_format.

response_formatCopy

{
  "type": "json_schema",
  "json_schema": {
    "name": "triage",
    "schema": {
      "type": "object",
      "properties": {
        "severity": {"type": "string", "enum": ["critical", "high", "medium", "low", "info"]},
        "reason": {"type": "string"}
      },
      "required": ["severity", "reason"]
    }
  }
}

Images

Send images to nebula-v as image_url content parts (an HTTPS URL or a data: URL).

Image content partCopy

{
  "model": "nebula-v",
  "messages": [{
    "role": "user",
    "content": [
      {"type": "text", "text": "What does this error page reveal?"},
      {"type": "image_url", "image_url": {"url": "data:image/png;base64,iVBORw0KGgo..."}}
    ]
  }]
}

Response

chat.completionCopy

{
  "id": "chatcmpl-...",
  "object": "chat.completion",
  "created": 1762862400,
  "model": "nebula-4.5",
  "choices": [{
    "index": 0,
    "message": {"role": "assistant", "content": "Usually medium..."},
    "finish_reason": "stop"
  }],
  "usage": {"prompt_tokens": 41, "completion_tokens": 88, "total_tokens": 129}
}

Note

A 402 means the organization has no spendable credit or its spend cap is reached. See Billing and credits.