API Quickstart - SuperAI Flows

This guide shows you how to authenticate, run your first workflow, and retrieve results.

Process a document in 60 seconds

Run this single script to authenticate, create a flow, and process your first invoice:

# Replace with your credentials
EMAIL="your-email@example.com"
PASSWORD="your-password"

# Get anonymous key and authenticate
ANON_KEY=$(curl -s https://flows.super.ai/api/auth/anon-key | jq -r .anon_key)
TOKEN=$(curl -s -X POST "https://core.flows.super.ai/auth/v1/token?grant_type=password" \
  -H "Content-Type: application/json" \
  -H "apikey: $ANON_KEY" \
  -d "{\"email\":\"$EMAIL\",\"password\":\"$PASSWORD\"}" | jq -r .access_token)

# Create flow with full definition
FLOW_RESPONSE=$(curl -s -X POST https://flows.super.ai/api/flows \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
  "display_name": "Quick Start: Invoice Processing",
  "description": "Extract and validate invoice data",
  "version": 1,
  "definition": {
    "flow": {"created_at": "", "author_uuid": ""},
    "flow_configs": [{
      "config": {
        "tasks": [
          {
            "visual_properties": {"position": {"x": 0, "y": 359}},
            "name": "8a0883ff-e0b9-4c27-8dfe-8af18c7d3ea4",
            "task_executor_name": "receive_file",
            "task_title": "Receive File",
            "parameters": {}
          },
          {
            "visual_properties": {"position": {"x": 530, "y": 232}},
            "name": "d61f3643-52d9-4ff3-bee7-7a3e26ca013f",
            "task_executor_name": "doc_to_text",
            "task_title": "Convert Document to Text",
            "parameters": {
              "worker": "gpt-4.1",
              "mime_type": "{{8a0883ff-e0b9-4c27-8dfe-8af18c7d3ea4.mime_type}}",
              "document_url": "{{8a0883ff-e0b9-4c27-8dfe-8af18c7d3ea4.document_url}}",
              "doc_representation": "whitespace"
            }
          },
          {
            "visual_properties": {"position": {"x": 1060, "y": 0}},
            "name": "ce65d796-fabe-4b05-be0e-e58182fdae98",
            "task_executor_name": "text_to_structured",
            "task_title": "Extract Invoice Data",
            "parameters": {
              "fields": [
                {"id": "invoice_no", "title": "Invoice Number", "data_type": "string", "instructions": ""},
                {"id": "total", "title": "total", "data_type": "float", "instructions": "identify the total amount"},
                {"id": "account_owner", "title": "Account Owner", "data_type": "string", "instructions": "Identify the account owner"}
              ],
              "worker": "gpt-4.1",
              "input_data": "{{d61f3643-52d9-4ff3-bee7-7a3e26ca013f.page_contents}}",
              "document_url": "{{d61f3643-52d9-4ff3-bee7-7a3e26ca013f.document_url}}",
              "extract_bounding_boxes": true,
              "use_logprobs_confidence": true
            }
          },
          {
            "visual_properties": {"position": {"x": 1590, "y": 285}},
            "name": "2dad3105-b031-4681-9588-0709b487034c",
            "task_executor_name": "collection_validation",
            "task_title": "Validate Invoice Total",
            "parameters": {
              "validators": [{"name": "Not Empty Validation"}],
              "input_collection": "{{ce65d796-fabe-4b05-be0e-e58182fdae98.total}}"
            }
          }
        ]
      }
    }]
  }
}')

FLOW_ID=$(echo "$FLOW_RESPONSE" | jq -r .id)
echo "Flow created: $FLOW_ID"

# Execute the flow with a sample invoice
EXEC_RESPONSE=$(curl -s -X POST https://flows.super.ai/api/flow-executions \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d "{\"flow_id\":\"$FLOW_ID\",\"input\":{\"filename\":\"invoice-example.pdf\",\"mime_type\":\"application/pdf\",\"document_url\":\"https://cdn.super.ai/invoice-example.pdf\"}}")

EXEC_ID=$(echo "$EXEC_RESPONSE" | jq -r .id)
echo "Execution started: $EXEC_ID"
echo "View results: https://flows.super.ai/flows/$FLOW_ID?view=data"

What just happened?

Authenticated with your SuperAI account.
Created a 4-task invoice processing flow: Receive File → Convert to Text → Extract Invoice Data → Validate.
Started processing a sample invoice PDF.
Got back a flow ID and execution ID to track progress.

Check the dashboard link to watch your flow process the document. Got an error? Jump to Troubleshooting.

What are SuperAI Flows?

SuperAI Flows automate multi-step document processing pipelines. A typical flow:

Classifies documents (invoice, receipt, contract)
Extracts structured data based on document type
Validates extracted values
Delivers results via webhook or API

Flow definition example

Flows are defined in declarative YAML that’s both human-readable and type-safe. Here’s a complete document processing workflow:

tasks:
  - name: classify_document
    task_executor_name: document_classifier
    parameters:
      document_url: "{{input.document_url}}"
      document_types:
        - invoice
        - bank_statement
        - contract

  - name: extract_data
    task_executor_name: text_to_structured
    parameters:
      document_url: "{{classify_document.splits[0].document_url}}"
      fields:
        - id: invoice_number
          data_type: string
        - id: total_amount
          data_type: float
          instructions: "identify the total amount"

  - name: validate_results
    task_executor_name: collection_validation
    parameters:
      input_collection: "{{extract_data.total_amount}}"
      validators:
        - name: Not Empty Validation

Key features:

Task chaining — extract_data references output from classify_document using {{classify_document.splits[0].document_url}}
Type safety — Each field specifies a data_type (string, float, etc.)
Built-in executors — Use pre-built task executors like document_classifier and text_to_structured
Validation — Chain validation tasks to ensure data quality

The platform automatically resolves dependencies, handles retries, and manages data flow between tasks.

Prerequisites

A SuperAI account (sign up free)
Your login email and password
curl and jq installed (for testing examples)

Authentication

SuperAI uses JWT-based authentication. All API requests require a valid access token.

Get the anonymous key

This is a public endpoint — no auth required.

curl https://flows.super.ai/api/auth/anon-key

{ "anon_key": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9..." }

Exchange credentials for JWT tokens

curl -X POST 'https://core.flows.super.ai/auth/v1/token?grant_type=password' \
  -H 'Content-Type: application/json' \
  -H 'apikey: YOUR_ANON_KEY' \
  -d '{"email": "your-email@example.com", "password": "your-password"}'

{
  "access_token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9...",
  "refresh_token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9...",
  "expires_in": 3600,
  "token_type": "bearer"
}

Use the access token in every request

curl https://flows.super.ai/api/flows \
  -H "Authorization: Bearer YOUR_ACCESS_TOKEN"

Token management

Token type	Lifetime	When to use
Access token	1 hour	Include in every API request
Refresh token	30 days	Get new access tokens when expired

Tokens grant full account access — treat them like passwords. Store them in environment variables or secret managers, never in code, and always use HTTPS.

When your access token expires, use the refresh token to get a new one without re-authenticating:

curl -X POST 'https://core.flows.super.ai/auth/v1/token?grant_type=refresh_token' \
  -H 'Content-Type: application/json' \
  -H 'apikey: YOUR_ANON_KEY' \
  -d '{"refresh_token": "YOUR_REFRESH_TOKEN"}'

Run your first flow

Start a flow by creating a flow execution with your input data.

curl -X POST https://flows.super.ai/api/flow-executions \
  -H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "flow_id": "YOUR_FLOW_ID",
    "input": {
      "filename": "invoice-example.pdf",
      "mime_type": "application/pdf",
      "document_url": "https://cdn.super.ai/invoice-example.pdf"
    }
  }'

{
  "id": "df77dc65-909b-430e-bff1-9e9ecf315a80",
  "flow_id": "550e8400-e29b-41d4-a716-446655440000",
  "flow_version": 1,
  "status": "running",
  "input": {
    "filename": "invoice-example.pdf",
    "mime_type": "application/pdf",
    "document_url": "https://cdn.super.ai/invoice-example.pdf"
  },
  "organization_id": "660e8400-e29b-41d4-a716-446655440001",
  "created_at": "2024-01-15T10:30:00Z",
  "updated_at": "2024-01-15T10:30:00Z"
}

Key fields:

id — The flow execution ID (use this to check status and retrieve results)
status — Current execution state (running, completed, failed)
flow_version — Version of the flow used (remains constant even if the flow is updated)

The input object structure depends on your flow’s configuration. For file-based flows, provide filename, mime_type, and document_url. Other flows may have different input schemas — check the flow’s input_schema.

Check execution status

curl "https://flows.super.ai/api/flow-executions/df77dc65-909b-430e-bff1-9e9ecf315a80" \
  -H "Authorization: Bearer YOUR_ACCESS_TOKEN"

{
  "id": "df77dc65-909b-430e-bff1-9e9ecf315a80",
  "flow_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "completed",
  "created_at": "2024-01-15T10:30:00Z",
  "updated_at": "2024-01-15T10:32:45Z"
}

Retrieve results

Get task execution results for your flow:

curl "https://flows.super.ai/api/task-executions?flow_execution_id=df77dc65-909b-430e-bff1-9e9ecf315a80" \
  -H "Authorization: Bearer YOUR_ACCESS_TOKEN"

{
  "task_executions": [
    {
      "id": "e47f2923-1205-4bf2-bee2-2974ddb64aad",
      "task_name": "extract_invoice_data",
      "status": "completed",
      "output": {
        "invoice_number": "INV-0052465571",
        "total_amount": 2435.00,
        "currency": "USD",
        "vendor": "Acme Corp",
        "date": "2024-01-15"
      },
      "created_at": "2024-01-15T10:30:05Z",
      "updated_at": "2024-01-15T10:32:40Z"
    }
  ]
}

Common query parameters:

flow_execution_id — Filter by flow execution (required for most queries)
task_name — Filter by specific task name
task_execution_idx — Filter by execution attempt (0 = first attempt, 1 = first retry)

Webhooks (recommended)

Instead of polling for results, configure webhooks to receive real-time notifications when flows complete. The Webhook Notification task sends POST requests to your specified URL. Add a Webhook Notification task to your flow with these parameters:

Url (required) — Your webhook endpoint URL
Authorization Header Value (optional) — Auth token or API key for securing your webhook
Payload (required) — The data to send (can reference task outputs using {{task_name.field}} syntax)

Example payload:

{
  "event_type": "flow_execution.completed",
  "flow_id": "550e8400-e29b-41d4-a716-446655440000",
  "flow_execution_id": "df77dc65-909b-430e-bff1-9e9ecf315a80",
  "status": "completed",
  "timestamp": "2024-01-15T10:32:45Z",
  "data": {
    "task_outputs": {
      "extract_invoice_data": {
        "invoice_number": "INV-0052465571",
        "total_amount": 2435.00
      }
    }
  }
}

Your webhook endpoint should validate the authorization header (if provided via the x-webhook-token header), process the payload, and return a 2xx status code to acknowledge receipt.

Troubleshooting

Authentication failed / Invalid credentials

Verify your credentials work in the web UI first at flows.super.ai.
Check for special characters in your password; wrap it in single quotes to preserve them: export PASSWORD='your-password'.
Verify the anonymous key is valid — it should be a long JWT string.

Flow not found (404)

Verify the flow was created successfully (check the flow creation response).
List all flows you have access to: curl https://flows.super.ai/api/flows -H "Authorization: Bearer $TOKEN".
Confirm you’re using the correct UUID and are logged into the correct organization — flows are organization-specific.

Unauthorized (401) mid-session

Your token expired after 1 hour. Use the refresh token to obtain a new access token (see Authentication).

Bad Request (400) when creating an execution

Validate your JSON is well-formed (pipe it through jq).
For custom flows, check the required input schema: curl "https://flows.super.ai/api/flows/$FLOW_ID" -H "Authorization: Bearer $TOKEN" | jq .input_schema.

Execution stuck in 'running'

Some flows take minutes for complex documents. Use webhooks instead of polling, and inspect task executions to see which task is blocking:

curl "https://flows.super.ai/api/task-executions?flow_execution_id=$EXEC_ID" -H "Authorization: Bearer $TOKEN" | jq '.task_executions[] | {task_name, status}'

Rate limiting (429)

The API implements fair-use rate limiting. If you receive 429 (Too Many Requests), wait before retrying and implement exponential backoff.

Still stuck? Check status.super.ai, then contact support@super.ai with your flow execution ID and the request_id from any error responses.

Error handling

The API uses standard HTTP status codes:

Code	Meaning	Example
200	Success	Request completed successfully
201	Created	Flow execution created
400	Bad Request	Invalid input data or missing required fields
401	Unauthorized	Missing or invalid access token
403	Forbidden	Insufficient permissions for this resource
404	Not Found	Flow or execution ID doesn’t exist
422	Validation Error	Request body failed schema validation
500	Server Error	Internal error (retry with exponential backoff)

Errors follow a consistent format. See the full Error Codes reference.

{
  "error": {
    "code": "not_found",
    "message": "Flow not found"
  },
  "request_id": "01K8KACP7D2XFGHJ9KLM4NPQR8"
}

Next steps

Downloading files

Retrieve gs:// files returned in your flow outputs.

API Reference

Full endpoint documentation with schemas.

​Process a document in 60 seconds

​What are SuperAI Flows?

​Flow definition example

​Prerequisites

​Authentication

​Token management

​Run your first flow

​Check execution status

​Retrieve results

​Webhooks (recommended)

​Troubleshooting

​Error handling

​Next steps

Downloading files

API Reference

Process a document in 60 seconds

What are SuperAI Flows?

Flow definition example

Prerequisites

Authentication

Token management

Run your first flow

Check execution status

Retrieve results

Webhooks (recommended)

Troubleshooting

Error handling

Next steps