Integrations
Overview
Integrations connect OutDo to external data sources and services. They allow you to pull data from APIs, import files, and use AI to extract structured information from documents.
Each integration is scoped to a workspace and stores its type-specific configuration. Integrations are managed through the Admin panel's Integrations section.
Integration properties
| Property | Description |
|---|---|
| Name/Label | Display name for the integration |
| Type | HTTP, Excel, AI Struct, or Template |
| Workspace | Which workspace the integration belongs to |
| Shared | Whether the integration is accessible from other workspaces |
HTTP integration
The HTTP integration connects to REST APIs, fetches data, maps response fields to feature columns, and syncs the results.
Configuration
Endpoints — Define one or more API endpoints:
| Setting | Description |
|---|---|
| Method | GET, POST, PUT, or DELETE |
| URL | API endpoint URL (supports {{org.var}} and {{ws.var}} template syntax) |
| Headers | Custom HTTP headers (e.g., authentication tokens) |
| Variables | Input parameters for the API call |
| Response structure | Expected JSON structure of the response |
Data mappings — Map API response fields to feature columns:
| Setting | Description |
|---|---|
| API path | JSON path in the response (e.g., data.0.customer_name) |
| Target feature | Destination feature for the mapped data |
| Target column | Destination column within the feature |
| Transformations | Value transformations (regex, formulas) applied during mapping |
Linked Select resolution — Automatically match API values to existing records in linked features.
User workflow
- Configure the endpoint URL, method, headers, and variables.
- Test the endpoint with sample variable values — view the response as JSON or table preview.
- Set up data mapping from API response fields to feature columns.
- Execute the fetch — data lands in a staging area for review.
- Validate the mapped data.
- Sync to the target feature.
Excel import/export
The Excel integration imports data from Excel (.xlsx) and CSV files with column mapping, transformations, and preview.
Configuration
File settings:
| Setting | Description |
|---|---|
| CSV delimiter | Separator for CSV files (default: comma) |
| CSV encoding | Character encoding (default: UTF-8) |
| Has header | Whether the first row contains column names |
| Start row | Number of rows to skip before data begins |
| Sheet name | For Excel files with multiple sheets |
Column mappings:
| Setting | Description |
|---|---|
| Source column | Column name in the uploaded file |
| Target column | Destination column in the feature |
| Component type | Target component type (for cleanup rules) |
| Include | Toggle to skip a column |
| Transformations | Data transformations per column (trim, uppercase, lowercase, regex, etc.) |
| Fixed values | Static values added to every imported row |
User workflow
- Upload an Excel or CSV file using the file picker.
- Select sheet (if the file has multiple sheets).
- Preview the first 100 rows in a table display.
- Review the column mapping configuration.
- Execute the import with a progress tracker.
- Review results — success count and error count.
AI Struct
The AI Struct integration uses artificial intelligence to extract structured data from documents (PDFs, images). It supports multiple AI providers and configurable extraction schemas.
Configuration
AI provider:
| Setting | Description |
|---|---|
| Provider | Mistral, OpenAI, or Anthropic |
| Model | Model name (e.g., GPT-4o, Claude Sonnet) |
| API key | Authentication key for the AI service |
| Endpoint | Custom API endpoint (optional) |
| Max tokens | Maximum output length |
| Temperature | Response creativity (0 = deterministic, 1 = creative) |
Input:
| Setting | Description |
|---|---|
| Mode | File picker (user selects files) or Variable (automated) |
| OCR | Optional OCR pre-processing with providers: Google ML Kit, Google Cloud Vision, Azure |
| Max pages | Limit the number of pages processed per document |
| PDF render DPI | Image quality for PDF conversion (default: 150) |
Extraction:
| Setting | Description |
|---|---|
| System prompt | Instructions for the AI on what to extract |
| User prompt template | Message template with {{instructions}} placeholder |
| Schema fields | Define the expected structure of extracted data |
| Column mappings | Map extracted fields to feature columns |
Deduplication:
| Setting | Description |
|---|---|
| Enabled | Check for duplicate records before import |
| Strategy | How to handle duplicates |
| Matching fields | Which fields to compare for deduplication |
User workflow
- Select documents to process (file picker or batch queue).
- Configure the AI provider, model, and extraction schema.
- Set the prompt instructions describing what to extract.
- Process files — AI extracts structured data.
- Review extracted data in a staging view.
- Handle duplicates based on deduplication rules.
- Sync to the target feature.
Auto-matching and auto-fill
AI Struct supports intelligent auto-matching of extracted data against existing records, and auto-filling feature columns based on header and line item mappings from processed documents.
Templates
Integration Templates provide pre-configured integration patterns that can be loaded and customized. They serve as starting points for common integration scenarios, reducing setup time.