curl --request POST \
--url https://api.opper.ai/v3/compat/v1/systemone \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"model": "<string>",
"questions": {},
"state": "<string>"
}
'import requests
url = "https://api.opper.ai/v3/compat/v1/systemone"
payload = {
"model": "<string>",
"questions": {},
"state": "<string>"
}
headers = {
"Authorization": "Bearer <token>",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)
print(response.text)const options = {
method: 'POST',
headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
body: JSON.stringify({model: '<string>', questions: {}, state: '<string>'})
};
fetch('https://api.opper.ai/v3/compat/v1/systemone', options)
.then(res => res.json())
.then(res => console.log(res))
.catch(err => console.error(err));{
"answers": {},
"model": "<string>",
"usage": {
"input_tokens": 1,
"output_tokens": 1
}
}{
"error": {
"code": "<string>",
"message": "<string>",
"details": "<unknown>"
},
"meta": {}
}{
"error": {
"code": "<string>",
"message": "<string>",
"details": "<unknown>"
},
"meta": {}
}System One evaluation
System One-compatible synchronous evaluation. Send model, state and named typed questions (noul, choice, score). TypeSafe models also take images and box questions (preview upstream; format and limits may change): send the TypeSafe-Danger-Zone-Images: true header, which is forwarded to TypeSafe, and list each image once in assets, keyed by a placeholder written where the image belongs in the state, instructions or criteria. An asset is base64 (data_base64 with format png, jpg or jpeg) in a JSON body, or a file part in a multipart/form-data body whose first part, request, holds the JSON body and whose asset names the part in part. A box question takes instructions and no criteria, needs Jev 1.14 (typesafe/jev-preview or typesafe/jev-1.14.0), and answers with box edges top, bottom, left and right as fractions of the image and a confidence. Images or a box question without the header return 400; other models return 422 for them. Bodies over 50 MiB return 413; TypeSafe’s own limits (a 32 MiB body, image size, dimensions, pixels) return its 400 or 413. A full catalog ID, for example typesafe/jev-1.13.0, typesafe/jev-latest or greenpt/metis, pins that endpoint. A bare model name, for example clef, routes over that model’s pooled endpoints as a chat model name does: a rate-limited or unavailable endpoint is retried once and the call falls over to the next, under the project’s Comply allowlist. List available endpoints with GET /v3/models?type=evaluation. GreenPT Metis jointly evaluates 1–32 questions in one request; state is required but may be null, and instructions may be omitted or null. Its nullable descriptions, probabilities, confidence, score legends and native impact metadata are preserved. EmpirioLabs Aplomb (empiriolabs/aplomb-1) takes 1–128 questions and images, audio and video as media nodes anywhere in state: an object with type image, audio or video and either a public https:// url or base64 data with its mime. Other providers retain their own validation contracts. Returns native model, answers and token usage. Uses a project-scoped Opper API key, Comply policies and provider entitlements. Cost and generation ID are returned in headers. Streaming, chat generation and tool calling are unsupported.
curl --request POST \
--url https://api.opper.ai/v3/compat/v1/systemone \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '
{
"model": "<string>",
"questions": {},
"state": "<string>"
}
'import requests
url = "https://api.opper.ai/v3/compat/v1/systemone"
payload = {
"model": "<string>",
"questions": {},
"state": "<string>"
}
headers = {
"Authorization": "Bearer <token>",
"Content-Type": "application/json"
}
response = requests.post(url, json=payload, headers=headers)
print(response.text)const options = {
method: 'POST',
headers: {Authorization: 'Bearer <token>', 'Content-Type': 'application/json'},
body: JSON.stringify({model: '<string>', questions: {}, state: '<string>'})
};
fetch('https://api.opper.ai/v3/compat/v1/systemone', options)
.then(res => res.json())
.then(res => console.log(res))
.catch(err => console.error(err));{
"answers": {},
"model": "<string>",
"usage": {
"input_tokens": 1,
"output_tokens": 1
}
}{
"error": {
"code": "<string>",
"message": "<string>",
"details": "<unknown>"
},
"meta": {}
}{
"error": {
"code": "<string>",
"message": "<string>",
"details": "<unknown>"
},
"meta": {}
}Authorizations
API key authentication. Pass your API key as a Bearer token.
Headers
Function name for tracing and project-level guardrail function-scope filtering.
Parent span ID for distributed tracing context.
Comma-separated key:value usage-attribution tags (e.g. tenant:acme,project:demo, max 8). Recorded on the generation's billing/metrics rows; group spend by any key via GET /v2/analytics/usage?group_by=. Header-borne twin of the /v3/session URL prefix tags (which win per key when both are present); opper.-prefixed keys and session_id are reserved. Malformed values return 400.
TypeSafe's opt-in for images and box questions: true. Required when the request has assets or a box question, and forwarded to TypeSafe as sent.
Body
- Option 1
- Option 2
1Show child attributes
Show child attributes
TypeSafe models only, with the TypeSafe-Danger-Zone-Images: true header. Images keyed by a placeholder written where each belongs in the state, instructions or criteria (inside any string value, not object keys). Every asset must be referenced.
Show child attributes
Show child attributes