Skip to main content
POST
/
v1
/
chat
/
completions
from auriko import Client

client = Client()
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)
{
  "id": "chatcmpl-abc123",
  "object": "chat.completion",
  "created": 1745100000,
  "model": "gpt-4o-2024-08-06",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "The capital of France is Paris."
      },
      "logprobs": null,
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 14,
    "completion_tokens": 7,
    "total_tokens": 21,
    "prompt_tokens_details": {
      "cached_tokens": 0
    },
    "completion_tokens_details": {
      "reasoning_tokens": 0
    }
  },
  "system_fingerprint": "fp_d08c293973",
  "routing_metadata": {
    "provider": "openai",
    "provider_model_id": "gpt-4o-2024-08-06",
    "model_canonical": "gpt-4o",
    "routing_strategy": "cost-focus",
    "throughput_tps": 9.2,
    "cost": {
      "usd": 0.000105
    }
  }
}

Documentation Index

Fetch the complete documentation index at: https://docs.auriko.ai/llms.txt

Use this file to discover all available pages before exploring further.

Creates a model response for the given chat conversation. Auriko routes the request to the optimal provider based on your routing preferences (cost, latency, throughput, etc.).

​
Auriko Extensions

Beyond OpenAI compatibility, this endpoint supports:
  • Multi-model routing: Use models[] instead of model to route across multiple models
  • Routing options: Control provider selection with the routing object
  • Provider extensions: Pass provider-specific parameters with extensions
  • Cost transparency: Response includes routing_metadata with cost breakdown
All parameters, request/response schemas, and examples are auto-generated from the OpenAPI specification.

Authorizations

​
Authorization
string
header
required

API key authentication. Keys start with ak_ prefix. Example: Authorization: Bearer ak_live_xxxxxxxxxxxx

Body

application/json
​
model
string
required

Model to route to, required for single-model requests. Mutually exclusive with gateway.models; providing both returns 400.

Examples: gpt-4o, claude-sonnet-4-6, llama-3.3-70b-instruct

Example:

"gpt-4o"

​
messages
object[]
required

The messages to generate a completion for

Minimum array length: 1
​
temperature
number

Sampling temperature (0-2). Some providers restrict this value when reasoning is enabled.

Required range: 0 <= x <= 2
​
top_p
number

Nucleus sampling parameter. Some providers restrict this value when reasoning is enabled.

Required range: 0 <= x <= 1
​
max_tokens
integer

Maximum tokens to generate (legacy, use max_completion_tokens)

Required range: x >= 1
​
max_completion_tokens
integer

Maximum tokens to generate. Reasoning models (o1/o3) use this field instead of max_tokens.

Required range: x >= 1
​
reasoning_effort
enum<string>

Controls reasoning effort for supported models. Auriko translates this into each provider's native reasoning control.

See /guides/extensions-and-thinking for per-provider behavior.

Available options:
low,
medium,
high,
xhigh,
max,
off
​
stop

Stop sequences. Restrictions vary by provider and model.

​
presence_penalty
number

Presence penalty (-2 to 2). Not supported by all providers.

Required range: -2 <= x <= 2
​
frequency_penalty
number

Frequency penalty (-2 to 2). Not supported by all providers.

Required range: -2 <= x <= 2
​
logit_bias
object

Token logit bias. Not supported by all providers.

​
seed
integer

Random seed for reproducibility

​
top_k
integer

Top-K sampling. Restricted by some providers when reasoning is enabled. Supported by Anthropic, Google, and vLLM.

Required range: x >= 1
​
min_p
number

Min-P sampling. Supported by some vLLM providers.

Required range: 0 <= x <= 1
​
top_a
number

Top-A sampling. Supported by some vLLM providers.

Required range: 0 <= x <= 1
​
repetition_penalty
number

Repetition penalty. Supported by vLLM providers.

​
tools
object[]

Tools the model can call

​
tool_choice
Available options:
auto,
required,
none
​
parallel_tool_calls
boolean

Allow parallel tool calls

​
functions
object[]
deprecated

Deprecated. Use tools instead. Auto-converted.

​
function_call
deprecated

Deprecated. Use tool_choice instead. Auto-converted.

Available options:
none,
auto
​
response_format
object
​
stream
boolean
default:false

Enable streaming responses

​
stream_options
object
​
user
string

User identifier for abuse detection

​
n
integer
default:1

Number of completions to generate. Not supported by all providers.

Required range: 1 <= x <= 10
​
logprobs
boolean

Return log probabilities. Not supported by all providers.

​
top_logprobs
integer

Number of top logprobs to return. Requires logprobs support.

Required range: 0 <= x <= 20
​
web_search_options
object

Web search configuration. Supported by OpenAI.

​
verbosity
string

Output verbosity control. Supported by OpenAI.

​
prompt_cache_key
string

Prompt caching identifier. Supported by OpenAI.

​
safety_identifier
string

Safety policy identifier. Supported by OpenAI.

​
gateway
object

Auriko routing, metadata, and multi-model configuration. Omit for default single-model routing.

​
extensions
object

Auriko extensions for provider-specific passthrough.

For reasoning control, use the top-level reasoning_effort parameter instead of extensions.

Provider Passthrough

Pass provider-specific parameters directly:

  • anthropic: Anthropic-specific parameters
  • openai: OpenAI-specific parameters
  • google: Google/Gemini-specific parameters
  • deepseek: DeepSeek-specific parameters

Passthrough parameters are forwarded as-is to the target provider.

Response

Successful completion.

For streaming (stream: true), responses are Server-Sent Events. Each event is a ChatCompletionChunk. The final chunk has choices: [] (empty) and contains usage and routing_metadata. Stream ends with data: [DONE].

​
id
string
required

Unique completion identifier

​
object
string
required
Allowed value: "chat.completion"
​
created
integer
required

Unix timestamp of creation

​
model
string
required

Model used for completion

​
choices
object[]
required

Completion choices

​
usage
object

Token usage statistics. Present in virtually all responses. May be absent in rare cases where upstream provider does not report usage.

​
system_fingerprint
string

Backend fingerprint for reproducibility. Not all models include this field.

​
routing_metadata
object

Routing decision metadata included in successful responses. 10 STABLE fields (4 required + 6 optional) in the current public contract.

​
service_tier
string | null

The service tier used for processing the request. Present for OpenAI-routed models.