Skip to main content
POST
Create a model response (OpenAI Responses, stateless); `stream: true` answers with `response.*` server-sent events

Authorizations

Authorization
string
header
required

ak_… project API key (or x-api-key)

Body

application/json

Rejected with 400 unsupported_parameter: previous_response_id, conversation, background: true, built-in tools, include values other than the two listed, item references, and any field not listed here.

model
string
required

Public model ID, e.g. a1 (A-05)

Minimum string length: 1
Example:

"a1"

input
required
instructions
string | null
max_output_tokens
integer | null

1…the model's max_output_tokens

Required range: x >= 1
temperature
number | null
Required range: 0 <= x <= 2
top_p
number | null
Required range: 0 <= x <= 1
tools
object[] | null
Maximum array length: 128
tool_choice
Available options:
none,
auto,
required
parallel_tool_calls
boolean | null
max_tool_calls
integer | null
Required range: x >= 1
reasoning
object | null

effort must be one of the model's reasoning_efforts

text
object | null
truncation
enum<string> | null
Available options:
auto,
disabled,
null
prompt_cache_key
string | null
Maximum string length: 512
include
string[] | null
stream
boolean | null

Default false

include_reasoning
boolean | null

false strips reasoning items from the output

policy_ids
string[] | null

Antara extension: published policies to add for this request

Maximum array length: 20
Required string length: 1 - 120
background
boolean | null

Only false

store
boolean | null

Always treated as false; the response echoes store: false

metadata
object | null
user
string | null
Maximum string length: 512
safety_identifier
string | null
Maximum string length: 64
service_tier
string | null
client_metadata
object | null

Response

application/json when stream is false; otherwise text/event-stream: event: <type> + data: <ResponseStreamEvent>, ending with response.completed or response.incomplete (response.failed on a failure after the headers)

id
string
required
Pattern: ^resp_[0-9a-f]{32}$
Example:

"resp_0193a6c2e7b47c3d9f1e2a4b5c6d7e8f"

object
enum<string>
required
Available options:
response
created_at
integer
required

Unix seconds

Example:

1793750400

status
enum<string>
required
Available options:
in_progress,
completed,
incomplete,
failed
error
object | null
required
incomplete_details
object | null
required
instructions
string | null
required
max_output_tokens
integer | null
required
max_tool_calls
integer | null
required
model
string
required
Example:

"a1"

output
object[]
required
parallel_tool_calls
boolean
required
previous_response_id
null
required
reasoning
object
required
store
enum<boolean>
required
Available options:
false
temperature
number | null
required
text
object
required
tool_choice
required
Available options:
none,
auto,
required
tools
object[]
required
top_p
number | null
required
truncation
enum<string>
required
Available options:
auto,
disabled
usage
object | null
required
metadata
object
required
user
string | null
required