Orchestria Guide
No-Code Data Transformation

1 Create a flow

Creating a flow is done in three simple steps. First, you define the input and output formats, then you provide data samples, and finally you describe the desired transformation.

1Choose formats

Select your input data format (JSON, XML, CSV, text, or PDF) and the desired output format. If using CSV, you can specify the separator and indicate whether your data contains a header row.

Step 1: Choose input formats (JSON, XML, CSV, Text, PDF) and output

2Provide samples

Copy a real sample of your input data and show what the expected result should look like. These samples allow the AI to understand exactly what you want to achieve. You can also import a file to automatically generate these samples.

⚠️ This data will be sent to AI to generate transformation code. We recommend anonymizing your samples.

Input sample:

json Copy

{
  "commande_id": "CMD-12345",
  "date": "2025-01-15",
  "montant": 150.00,
  "client": {
    "nom": "Dupont",
    "email": "dupont@example.com"
  }
}

Output sample:

json Copy

{
  "id": "CMD-12345",
  "date_formatee": "15/01/2025",
  "total_ttc": 150.00,
  "client_email": "dupont@example.com"
}

Step 2: Provide samples and describe the desired transformation

2 Monitoring and Alerts

Define business alerts when creating the flow or via the AI chat, and receive automatic notifications.

Define alerts

You can define alerts when creating the flow or via the AI chat. For example: "Add an alert if the amount exceeds €10,000" or "Notify me if the status is error".

Email notifications

Enable the "Email notifications" option in the flow settings to receive an email for each alert or error.

3 AI Provider Choice

Orchestria supports multiple artificial intelligence providers. You can choose the one that best suits your needs in terms of quality and cost.

Available providers

How to change provider

Go to My Account → Settings to select your AI provider. The change applies immediately to all your operations: flow creation, AI chat modifications, and PDF extraction.

4 PDF Flows

Orchestria automatically extracts structured data from your PDF documents (invoices, quotes, purchase orders...) using AI Vision.

AI Vision Extraction

Each PDF is analyzed by the vision AI (based on the provider configured in your settings) that identifies and extracts all relevant fields: numbers, dates, amounts, invoice lines, etc.

The 3 validation modes

When creating a PDF flow, you choose how documents will be processed at execution:

PDF Validation Page

Access the "PDF Validations" page from the menu to manage pending documents:

• PDF preview and extracted data side by side
• Confidence score and validation rules details
• Edit extracted JSON before validation
• Bulk actions (validate/reject multiple PDFs)

JSON in PDF flows

When creating the flow

When you create a PDF flow, the AI analyzes your sample document and automatically generates a JSON representing the data structure. This JSON becomes a template that will be used for all future PDFs sent to this flow.

This template defines:

• The data structure the AI will extract from each PDF
• Available fields (number, date, amounts, line items...)
• Sample values (which will change with each document)

{
  "numero_facture": "F-2024-001",
  "date": "2024-01-15",
  "fournisseur": "Boulanger",
  "total_ttc": 219.99,
  "lignes": [
    { "designation": "Produit A", "quantite": 1, "prix_ht": 183.33 }
  ]
}

What you can modify

The generated JSON is fully customizable:

• Rename fields to match your system
• Remove fields you don't need
• Add fields if you know they exist in your documents
• Apply a predefined template (Invoice, Order, Quote, Delivery Note) to start with a structure suited to your document type

At execution

When you send a PDF to the flow, the AI uses the template defined at creation to extract data. It fills each field of the template with values found in the document.

The process:

1. AI reads the PDF and extracts data
2. A confidence score is calculated (amount consistency, required fields...)
3. Based on validation mode, the JSON is auto-validated or queued for review
4. You can correct extraction errors before validation
5. Once validated, the JSON is transformed to the output format

From input to output

The extracted JSON (input) is transformed by your flow to the output format. The prompt you wrote describes this transformation.

Executing PDF flows

Once the flow is created, you can execute it via the web interface or via API by sending the PDF as binary.

API call example

curl -X POST https://orchestria.io/api/flows/YOUR_FLOW_UID/your-flow-name \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/pdf" \
  -H "X-Filename: invoice.pdf" \
  --data-binary @invoice.pdf

5 Test and execute

Once your flow is generated, you access an interface divided into two parts: execution on the left, chat with the AI agent on the right.

Execution interface with input form on the left and results on the right

Execute the flow

Paste your real data into the input field, verify that the format matches the configured one, then click "Execute". The result is displayed immediately with the chosen output format. You can download the result or copy it to your clipboard.

Format management

The result is displayed in the output format configured when creating the flow. For asynchronous webhooks, the system returns an execution identifier to track processing in the logs.

Results displayed with configured output format and download option

6 AI Agent

The AI agent allows you to refine your flow through successive iterations, directly via chat. You describe what you want to modify, and the AI adjusts the code accordingly.

AI chat to modify the flow through successive iterations

Use the chat

Simply describe the desired modifications. For example: "Add a filter to exclude inactive customers" or "Change the date format to MM/DD/YYYY". The AI understands your flow's context and proposes an adapted modification. Be specific in your requests and mention the concerned field names.

Code validation system

When the AI modifies your flow's code, the new code is automatically deployed so you can test it. A validation banner appears above the chat, temporarily blocking new modifications. Test the flow with the execution form on the left, then click "Validate" if the result suits you, or "Rollback" to return to the previous version.

Automatic rollback

If you choose rollback, the system automatically restores the old code version and redeploys it. A system message appears in the chat to confirm the rollback. This feature ensures you never lose a working version of your flow.

7 Deploy to production

Once your flow is tested and validated, you can integrate it into your applications via a secure REST API.

Get the execution URL

Click the "eye" icon next to your flow in the list to display its execution URL. This URL is unique and permanent. You can copy it and use it in your applications. Calls are made via POST with your data in the request body.

Generate an API key

Go to the "API" section in the menu to generate an authentication key. This key must be transmitted in the Authorization: Bearer YOUR_KEY header for each call. Never share this key publicly, it gives access to all your flows.

API key management page to authenticate your calls

📘 API call examples

curl -X POST https://orchestria.io/api/execute/YOUR_FLOW_ID \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "commande_id": "CMD-12345",
    "montant": 150.00
  }'

import requests

url = "https://orchestria.io/api/execute/YOUR_FLOW_ID"
headers = {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json"
}
payload = {
    "commande_id": "CMD-12345",
    "montant": 150.00
}

response = requests.post(url, json=payload, headers=headers)
result = response.json()
print(result)

const response = await fetch('https://orchestria.io/api/execute/YOUR_FLOW_ID', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer YOUR_API_KEY',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    commande_id: 'CMD-12345',
    montant: 150.00
  })
});

const result = await response.json();
console.log(result);

8 Connectors

Connectors allow you to fully automate your integrations by retrieving data from external sources (polling) or sending your results to third-party services (output webhooks).

Connector configuration interface (input and output)

Input Connectors (HTTP Polling & SFTP)

Input connectors automatically retrieve data from external sources.

As soon as new data is detected, the connector automatically triggers the flow configured in the Routing page. Polling is counted in your quota even if no data is retrieved, so choose a frequency adapted to your real needs.

Output Connectors (Webhooks & SFTP)

Output connectors automatically send your flow results to an external destination.

Advantage: Create a connector once and reuse it across multiple flows (instead of configuring webhooks individually for each flow in the Routing page).

Configuration and security

Connectors support four authentication methods: none, Bearer Token, Basic Auth, and Sign-In (for APIs requiring prior login generating a JWT token or API key). For input connectors, the minimum polling frequency depends on your plan (60min for Free, 30min for Starter, 10min for Pro, 5min for Enterprise). All connectors benefit from detailed logs and automatic error handling for maximum reliability.

9 Advanced features

Flow routing and chaining

The routing system allows you to automatically chain multiple flows. The first flow's result becomes the second's input, and so on. You only receive the first flow's result in response, the following ones execute in the background. To retrieve the final result of a chain, configure a webhook on the last flow in the sequence.

Visualization of flow chaining with interactive diagram

Virtual DNS for multiple environments

Virtual DNS allows you to create distinct deployment environments (production, staging, development) with custom URLs. You can test a new version of your flow on a test URL, then switch the production URL to this new version once validated. This enables zero-downtime deployment and instant rollback if needed.

Logs and diagnostics

The "My Logs" section centralizes all your flow executions. You can filter by status (success, error, running, rejected) and view details of each execution. Pro+ users benefit from automatic AI error analysis that precisely identifies problems (incorrect format, missing field) and suggests correction paths.

10 Security and Privacy

Protecting your data is our absolute priority. Orchestria is designed to ensure total confidentiality of your business data.

AI and data separation

Artificial intelligence is only used during flow creation to generate transformation code. It only has access to the samples you provide, never to your real production data. Your execution payloads never transit through AI services.

No data storage

Data you send for execution is processed in memory then immediately deleted. Orchestria stores no payloads, which makes execution replay impossible but guarantees total GDPR compliance. Only metadata (status, timestamp, user) is retained for logs.

Sample protection

The data samples you provide when creating a flow are securely stored in your account and never shared with third parties. These samples are necessary for the flow to function and can be modified by editing the flow.

AI and PDF documents

When creating or editing a PDF flow, the document is sent to the AI for processing. Only "PDF Vision" mode uses AI for data extraction at each execution. This mode consumes more resources and should be used knowingly for complex or scanned documents.

11 Quotas

Each plan has monthly quotas. Here are the different types of quotas.

Number of flows

Limit on the number of flows you can create. Varies by plan (e.g., 5 flows for Starter, unlimited for Enterprise).

Executions

Monthly API calls. Each call to a flow counts as an execution, whether it succeeds or fails.

AI Credits

Used when creating, modifying, and improving flows via the AI chat. PDF Vision mode also consumes AI credits.

Check your quotas

View your consumption on the "My Account" page or on the dashboard.

12 Technical limits

To ensure optimal performance and service stability for everyone, Orchestria applies certain technical limits to processed payloads.

Maximum payload size

The maximum request size (input payload) is 10 MB. This limit applies uniformly to all formats (JSON, XML, CSV, text) and all execution methods (REST API, webhooks, connectors). Beyond this limit, you will receive a 413 error (Request Entity Too Large).

Capacity by format

A 10 MB limit allows processing approximately 500,000 simple JSON objects, 100,000 to 200,000 CSV rows, or 20,000 complex XML elements. This capacity covers 95 to 98% of standard integration use cases. For larger volumes, we recommend splitting your data into multiple calls or contacting us to discuss a solution adapted to your needs.