Platform guide

Create a receipt from StackAI

Use a StackAI action or custom API-style step to record what an agent or workflow did.

Agent workflow

Add the receipt step after the agent output

Create the receipt after StackAI has generated, searched, classified, drafted, or acted, so the payload can describe the work in practical language.

Use the Automation Receipts API endpoint

Send create-run requests to https://automationreceipts.com/api/v1/runs. If your account uses a different workspace or environment endpoint, use the endpoint shown in your Automation Receipts setup.

Use the StackAI action payload

Starter file

Download this JSON body for an action/custom API-style step, then map node outputs into run_uid, summaries, tools_used, and final_action.

Download StackAI payload JSON

What you need

Automation Receipts access

Use an active account and an active API key. The key is sent as a bearer token.

StackAI workflow

Start with a controlled workflow run and safe example values before connecting private data, customer-facing actions, or live approval steps.

Basic workflow shape

Put the receipt step after the workflow has enough information to summarize the trigger, output, review state, and final action.

StackAI input
  ->
LLM, knowledge, or app nodes
  ->
Action / Custom API step
  ->
Receipt appears in Automation Receipts

Request settings

Step: Action or Custom API-style request
Method: POST
URL: https://automationreceipts.com/api/v1/runs
Content type: application/json
Headers: Authorization and Content-Type
Body: JSON payload built from node outputs

Headers

Authorization: Bearer aar_live_REPLACE_ME
Content-Type: application/json

Sample payload

Use the platform's mapped values where the example says to replace a value. Keep summaries readable and avoid sending raw secrets, full customer messages, or unnecessary private payloads.

Create-run JSON

{
  "automation_name": "StackAI Policy Assistant",
  "run_uid": "stackai-REPLACE_WITH_RUN_ID",
  "source_type": "stackai",
  "trigger_type": "agent_run",
  "status": "needs_review",
  "risk_level": "medium",
  "input_summary": "A teammate asked the agent to summarize a policy question.",
  "output_summary": "StackAI generated an answer using the configured knowledge sources.",
  "tools_used": "StackAI workflow, knowledge base, LLM",
  "approval_required": true,
  "approval_status": "pending",
  "final_action": "Answer saved for review before sharing.",
  "events": [
    {
      "event_type": "data_accessed",
      "event_label": "Knowledge checked",
      "event_summary": "The workflow searched the configured knowledge source."
    },
    {
      "event_type": "output_generated",
      "event_label": "Answer generated",
      "event_summary": "The agent generated a draft answer for review."
    }
  ]
}

Make `run_uid` unique

Map a StackAI run ID, conversation/message ID, timestamp, or another unique workflow value into run_uid. Automation Receipts rejects duplicate run_uid values within the same workspace.

Optional timeline events

For the first controlled run, include a short events array in the create-run payload. If the workflow continues later, append an event with the original run_uid.

Append an event later

POST https://automationreceipts.com/api/v1/runs/stackai-sample-run-001/eventsAuthorization: Bearer aar_live_REPLACE_ME
Content-Type: application/json

{
  "event_type": "action_taken",
  "event_label": "Workflow finished",
  "event_summary": "The workflow finished and the receipt was updated."
}

StackAI notes

Starter file scope

This download is a copy-ready action payload. It is not a full StackAI project export or marketplace template.

Record the agent result, not every token

Receipts should summarize what the agent did and what happened next, not store complete prompts, private documents, or full model output unless needed.

Use node outputs deliberately

Map only the fields that help a person understand the run: input summary, output summary, tools used, review status, and final action.

Review is optional

Use approval_required true only when someone needs to record a review outcome. Otherwise use approval_status not_required.

Start with a manual test

Run the agent with safe example data before wiring customer-facing actions or internal knowledge sources.

Troubleshooting

Problem	Likely cause	What to check
401 unauthorized	The bearer token is missing, revoked, or not the full API key.	Use the complete key value shown when it was created, prefixed with Bearer.
409 duplicate run_uid	A receipt already exists with that run ID in the workspace.	Use a dynamic value that changes for each workflow run.
422 validation error	A required field is missing or a machine value is not allowed.	Check automation_name, run_uid, summaries, source_type, status, and approval_status.

Next steps

StackAI agent recipe

Record an agent answer, knowledge use, and optional review outcome.

Payload fields

Check field meanings, allowed values, request limits, and event types.

Recipe library

Browse receipt patterns for other platforms and workflows.

StackAI Action Node docs

Open the platform documentation for the HTTP, webhook, or action step used in this guide.