Workspace AI
Connect Chat, Library, Dashboard, API, and integrations with company systems.
Integrate Workspace AI capabilities into your company systems with authentication, structured responses, traceability, and approved-access flows.
The Apeirum API is the technical path for connecting parts of Workspace AI to digital products, internal systems, and business routines. Access is approved by scope, and final integration contracts are defined with the Apeirum team.
Connect Chat, Library, Dashboard, API, and integrations with company systems.
Return answer, structured_data, evidence, status, job_id, and metadata when applicable.
Credentials, limits, schemas, and support are defined before production.
The API is not an unrestricted public endpoint. Usage depends on approval, commercial agreement, and technical alignment.
The integration pattern uses an API key in a header. In production, treat the key as a server-side secret and never expose it in a client.
The examples below show the expected integration format. Base URL, endpoints, and final schemas may vary according to the approved contract.
curl -X POST "$APEIRUM_BASE_URL/v1/answers" \
-H "x-api-key: $APEIRUM_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"question": "Quais itens exigem revisão?",
"scope": "workspace_contract",
"include_evidence": true
}'const response = await fetch(
`${process.env.APEIRUM_BASE_URL}/v1/answers`,
{
method: "POST",
headers: {
"x-api-key": process.env.APEIRUM_API_KEY,
"Content-Type": "application/json"
},
body: JSON.stringify({
question: "Quais itens exigem revisão?",
scope: "workspace_contract",
include_evidence: true
})
}
);
const data = await response.json();$headers = @{
"x-api-key" = $env:APEIRUM_API_KEY
"Content-Type" = "application/json"
}
$body = @{
question = "Quais itens exigem revisão?"
scope = "workspace_contract"
include_evidence = $true
} | ConvertTo-Json
Invoke-RestMethod `
-Method POST `
-Uri "$env:APEIRUM_BASE_URL/v1/answers" `
-Headers $headers `
-Body $bodyA typical integration sends a document or data, queries authorized context, receives a structured response, and tracks status when processing is asynchronous.
Document, data, event, or question arrives with a defined purpose.
The integration informs scope, user, routine, or authorized collection.
The API returns answer, structured_data, evidence, status, and metadata when applicable.
Asynchronous flows expose job_id and status for tracking.
When contracted, webhooks return relevant events to the destination system.
The paths below are scoped integration contract examples. They document the expected design without promising unrestricted public availability.
Model
Base URL, version, and final payload are defined in the approved scope.
Send a document or file for approved processing.
Check status, error, or completion of asynchronous processing.
Send a question or data and receive a structured response.
Register a callback when the contracted flow requires asynchronous return.
Responses may combine text, structured data, evidence, and metadata. Final fields depend on the contracted schema.
{
"job_id": "job_01hx_example",
"status": "completed",
"answer": "Há três itens que exigem revisão antes de seguir.",
"structured_data": {
"items": [
{
"label": "Prazo de renovação",
"risk": "medium",
"requires_review": true
}
]
},
"evidence": [
{
"document_id": "doc_01hx_example",
"excerpt": "Renovação automática por 12 meses salvo aviso prévio.",
"location": "page 4"
}
],
"metadata": {
"request_id": "req_01hx_example",
"scope": "workspace_contract"
}
}The codes below guide failure handling in clients and integrations. Final messages may vary by contract.
Fix payload, required fields, or format.
Check missing or invalid API key.
Confirm whether the contracted scope permits the action.
Check identifier, job, or requested resource.
Respect rate limit and retry with backoff.
Log request_id and contact technical support if it persists.
Limits are defined by scope, plan, or contract. Billing may consider plan, volume, authorized endpoints, support, and request usage.
The integration should preserve operational control. Sensitive cases need logs, permissions, human review, and limits before external actions.
Talk about destination system, volume, data, expected responses, callbacks, and review requirements. Apeirum defines the technical contract from the approved scope.