AuraOne Developer Home
Explore full API docs, SDK packages, signed webhook guides, and technical references across AuraOne.
Read the API docsENTERPRISE INTELLIGENCE API
Define work, start runs, route exceptions, receive decisions, and inspect what happened. Bring Enterprise Intelligence into your own systems through versioned interfaces.
The Enterprise Intelligence API provides versioned /api/v1/enterprise-intelligence/* endpoints, typed TypeScript and Python SDKs, and HMAC-SHA256 signed webhooks for triggering workflow runs, tracking human reviews, and retrieving evidence-backed deliveries.
01 / Versioned request
Create or reference an approved workflow and initiate a run from your existing application. All keys below are redacted placeholders—never real credentials.
curl -X POST https://api.auraone.ai/api/v1/enterprise-intelligence/runs \
-H "Authorization: Bearer ${AURAONE_API_KEY}" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: run-2026-11-20-batch-14" \
-d '{
"workflowVersionId": "wfv_example_v3",
"inputContentId": "cnt_example_batch_14",
"environmentVersion": "env-2026.11"
}'
# 202 Accepted
# {
# "data": { "id": "run_example_01", "status": "QUEUED" },
# "links": { "self": ".../runs/run_example_01" }
# }Auth, scope & error contract
Every route except the unauthenticated health probe binds organization scope from the credential and returns structured JSON errors.
02 / State & provenance
Track state, responsible reviewer, exceptions, and acceptance as the workflow progresses.
01
POST /api/v1/enterprise-intelligence/runs returns 202 with a run ID and polls cleanly through QUEUED and RUNNING states.
GET /runs/{runId}
02
When policy uncertainty triggers human review, state transitions to WAITING_FOR_HUMAN and surfaces the assigned review task.
GET /human-tasks
03
Inspect the output payload, execution trace, lineage, and sealed delivery manifest once checks and reviews settle.
GET /runs/{runId}/delivery
04
Verify the HMAC-SHA256 signature over timestamp + raw request bytes before applying the accepted result in your downstream system.
X-Webhook-Signature
03 / Signed webhooks
Use signed webhooks and documented verification to connect the decision to your own system. Verify the HMAC-SHA256 signature over ${timestamp}.${rawBody} against the raw request bytes before parsing JSON.
import { createHmac, timingSafeEqual } from "node:crypto";
const TOLERANCE_MS = 5 * 60 * 1000;
export function verifyAuraOneWebhook(
rawBody: Buffer,
headers: Record<string, string | string[] | undefined>,
secret: string,
now = Date.now(),
): boolean {
const signature = String(headers["x-webhook-signature"] ?? "");
const timestamp = String(headers["x-webhook-timestamp"] ?? "");
if (!secret || !signature || !/^\d+$/.test(timestamp)) return false;
const sentAt = timestamp.length > 11 ? Number(timestamp) : Number(timestamp) * 1000;
if (Math.abs(now - sentAt) > TOLERANCE_MS) return false;
const provided = signature.startsWith("sha256=") ? signature.slice(7) : signature;
if (!/^[0-9a-f]{64}$/i.test(provided)) return false;
const expected = createHmac("sha256", secret)
.update(`${timestamp}.`)
.update(rawBody)
.digest();
const received = Buffer.from(provided, "hex");
return received.length === expected.length && timingSafeEqual(received, expected);
}04 / Typed SDKs
Both @auraone/sdk (TypeScript) and aura_one (Python) target the same /api/v1/enterprise-intelligence routes, attach idempotency-key headers automatically on mutations, and raise explicit typed errors on 503 or permission denials.
import { AuraOneClient } from "@auraone/sdk";
const client = new AuraOneClient({
apiKey: process.env.AURAONE_API_KEY!, // ak_00000000_redacted...
baseUrl: "https://api.auraone.ai",
});
// Start a workflow run with an explicit idempotency key
const created = await client.enterpriseIntelligence.createRun({
workflowVersionId: "wfv_example_v3",
inputContentId: "cnt_example_batch_14",
environmentVersion: "env-2026.11",
idempotencyKey: "run-2026-11-20-batch-14",
});
// Inspect status, trace, and sealed delivery evidence
const run = await client.enterpriseIntelligence.getRun(created.data.id);
const delivery = await client.enterpriseIntelligence.getRunDelivery(created.data.id);05 / Connector scope
See which connectors are supported in the current adapter catalog, which read or write operations each one permits, and what authorization is required before production writes run. Every connector requires customer-specific registration and authorization rather than blanket activation.
Every write needs a write-scope grant, a human approval and an idempotency key. In production, a write also needs a verified sandbox proof.
| Connector | Read | Write |
|---|---|---|
| HubSpot CRM | Read contacts | Upsert contacts |
| Salesforce CRM | Read contacts | Upsert contacts |
| Jira | None | Create issue |
| Slack | None | Post run summary |
| Microsoft Teams | None | Post card |
| Amazon S3 | Check bucket access | Write object |
| Snowflake | None | Provision warehouse connection |
| BigQuery | None | Stream rows (start sync) |
| Databricks | None | Run job |
| PagerDuty | None | Trigger incident |
Full route inventory
Every family below corresponds directly to the live /api/v1/enterprise-intelligence route tree. Mutating handlers write audit events automatically.
| API family | Covers | Surface |
|---|---|---|
/api/v1/enterprise-intelligence/workflows | Workflow definitions, versions, activations, approval requests, per-workflow evaluations, release gates, chart review, the weekly queue, and executive buyer reports (/workflows/{workflowId}/buyer-report) | Workflow scoping and versioning |
/api/v1/enterprise-intelligence/runs | Runs plus per-run acceptance, auditor export, cancel, commercial terms, compensations, delivery, derivatives, elapsed time, incidents and human incident severity triage (/runs/{runId}/incidents/{incidentId}/triage), run lineage (/runs/{runId}/lineage), output, field queries, replay, and trace | Run lifecycle and evidence |
/api/v1/enterprise-intelligence/deliveries | Accepted deliveries and per-delivery detail | Delivery receipts |
/api/v1/enterprise-intelligence/batches | Batch ingest, items, acceptance, auditor export, cancel, connector delivery, and retry | Batch intake |
/api/v1/enterprise-intelligence/human-tasks | Review tasks, completion, and finding decisions for cases that pause for qualified review | Human review |
/api/v1/enterprise-intelligence/connectors | Connector registration, authorization, inbound endpoints, pull, receipts, test, ticket deliveries, and revocation | Integrated systems |
/api/v1/enterprise-intelligence/tool-bindings | Tool bindings attached to approved workflow versions | Bound actions |
/api/v1/enterprise-intelligence/evaluations | Evaluation records that judge workflow versions and candidates | Evaluation |
/api/v1/enterprise-intelligence/improvement-items | Named failures and improvement targets with owner and status | Model improvement |
/api/v1/enterprise-intelligence/regression-suites | Regression suites, suite runs, and exports | Regression control |
/api/v1/enterprise-intelligence/historical-sets | Historical task sets: register, examples, validation, baseline, dataset, acceptance | Evidence data |
/api/v1/enterprise-intelligence/context-sources | Context source registration, ingest, retry, and withdraw | Permissioned context |
/api/v1/enterprise-intelligence/registry | Model, prompt, tool, node-option, and policy registries | Scoped registries |
/api/v1/enterprise-intelligence/operations | Operations metrics, automation trend, SLOs, SLO reports, and the SLO release gate | Operations |
/api/v1/enterprise-intelligence/settings | Per-organization policy versions, the ticketing switch, and the connector sandbox-proof gate | Administration |
/api/v1/enterprise-intelligence/run-inputs | Eligible input content references and environment versions for a new run | Run intake |
/api/v1/enterprise-intelligence/menu | Menu operations and menu photo readings with their proposals | Menu workflow |
/api/v1/enterprise-intelligence/health | Unauthenticated probe: worker heartbeat and dependency status; 503 when unavailable | Health |
Additional domain-specific and operational endpoints under the /api/v1/enterprise-intelligence prefix.
| Route directory | Covers |
|---|---|
claims/ | Claims operations, operation previews, and per-run claim decisions |
compensations/ | Execute one compensating action raised on a run |
dashboard/ | Organization summary of workflows, runs, review queue, usage, invoices, and margins |
improvements/ | Improvement cycles and their human-approved state transitions |
regression-candidates/ | Propose a regression case from an incident and review it |
workflow-schedules/ | Workflow schedules and their latest cycles |
What comes next
Explore full API docs, SDK packages, signed webhook guides, and technical references across AuraOne.
Read the API docsSee how scoped connectors, human review pauses, and destination writes operate in connected enterprise workflows.
See Integrated workflowsScope an enterprise workflow and provision organization-bound API keys with your engineering team.
Request API accessCommon questions
Keys use the ak_<8 hex>_<64 hex> format passed via Authorization: Bearer or x-api-key. Your organization is derived strictly from the key record; inbound X-Org-Id headers are scrubbed by middleware. Missing or insufficient scopes return structured 401 or 403 JSON errors ({ error: { code, message } }).
State-changing routes such as POST /api/v1/enterprise-intelligence/runs require an idempotency-key header (8–255 characters). Replaying the same key returns the committed outcome envelope with x-idempotency-replayed: true.
No. Each supported connector requires explicit registration, scoped authorization, and customer configuration inside your engagement. Connectors not listed below are not claimed as generally supported.