February 2026
February 25, 2026
- PaddleOCR Engine: Added PaddleOCR as a new OCR engine with an updated model for Latin languages and memory-efficient model loading
- LLM Cost Tracking: Billing transactions now include per-extraction LLM token counts and costs for full cost transparency
- Table Extraction Accuracy: Fixed cross-row bounding box contamination with Y-coordinate constraints and improved prompts for tables containing surrounding text
- Extraction Pipeline Metrics: New Prometheus instrumentation and Grafana dashboards for real-time visibility into extraction pipeline performance and worker queues
- Auto-scaling Workers: Extraction workers now auto-scale based on queue depth using KEDA and Prometheus metrics
- Sample Workflows: New accounts are prepopulated with sample workflows and documents for a guided onboarding experience
- Langfuse Tracing: Improved extraction tracing with workflow names, LLM generation spans, and decoupled OpenTelemetry integration
- Bug Fix: Fixed a crash in smart lookup when the search corpus was empty
- Bug Fix: Fixed N+1 queries on extraction callbacks and field list endpoints
- Bug Fix: Fixed file download filenames not being readable by the frontend
February 21, 2026
- Visual Grounding 2.0: Completely revamped evidence highlighting with block-level and element-level precision, making it easier to trace extracted values back to their exact location in the original document
- Billing Dashboard: New usage dashboard with remaining credits, weekly summaries, per-workflow usage breakdown, and CSV export
- Document Segmentation: Advanced document layout analysis now enabled in production for improved table and section detection
- Bug Fix: Fixed an issue where bounding box highlights could remain stuck after navigating between fields
- Bug Fix: Fixed calibration configuration for confidence scoring
February 14, 2026
- Redesigned Home Page: New personalized home with recent workflows, at-a-glance file status indicators, and a streamlined workflow creation experience with templates
- New Schema Builder: Redesigned workflow configuration with a two-step Define & Refine flow, AI-powered field suggestions, inline file previews, and drag-and-drop uploads
- Faster Page Rotation: Upgraded page orientation detection engine for ~2x faster processing per page
- Improved Field Suggestions: More reliable AI field suggestions powered by an upgraded model
- Smart Lookup Fix: Fixed an issue where smart lookup settings were not saved when publishing a workflow version
- Bug Fix: Fixed sidebar overflow in the workflow studio
February 7, 2026
- New Dashboard Design: Refreshed visual design across the platform with updated components, colors, and layout
- Improved Visual Grounding: Evidence highlighting now uses an event-driven architecture, making interactions faster and more reliable across all views
- Secure File Access: Files and organization logos are now served securely through the backend, improving security and simplifying local development
- Workflow Name Suggestions: New AI-powered workflow name suggestions based on your description and uploaded files
- File Status Counts: Workflow listings now show at-a-glance counts of files by processing status
- Bug Fix: Fixed document element classification that could cause incorrect layout detection in certain documents
February 3, 2026
- Smart Lookup: New feature to enrich extracted data with external reference files using AI-powered matching
- Field Source Tracking: Fields now track their source (extraction vs. smart lookup) for better traceability
- OCR Performance: Improved startup times for OCR services with memory snapshots
- Bug Fix: Fixed field data processing to correctly handle manual field creation
January 2026
January 27, 2026
- Usage-Based Billing: Introduced billing for data extraction operations with transparent per-page pricing
- Improved Suggestions: Faster and more reliable field suggestions when creating workflows
- Bug Fix: Resolved an issue where certain field value formats could cause upload errors
January 20, 2026
- Faster OCR Processing: Improved document processing speed with optimized text recognition
- Extraction Review: Administrators can now review extractions directly in the dashboard
- Organization Credits: Added ability to manage organization credits for usage billing
January 13, 2026
- Faster Document Processing: Significantly improved processing speed for large documents
- Row Count Filters: Filter extraction results by the number of rows extracted
- Page Rotation Control: Configure custom page rotation strategies per workflow
January 6, 2026
- Results Filtering: Enhanced results download with additional filtering options
- API Documentation: Expanded API reference documentation with more examples
December 2025
December 15, 2025
- Improved Reliability: Enhanced system stability and faster deployments
December 1, 2025
- Performance Improvements: Optimized results retrieval for workflows with many documents
- Cost Tracking: Added detailed usage tracking for extraction operations
November 2025
November 24, 2025
- Datarow Creation: Create new data rows directly from specific extractions
November 17, 2025
- Table Extraction: Improved accuracy when extracting data from complex tables
- User Status Endpoint: New
/user/is-enabledendpoint to check account status - File Validation: File validator responses now include user information
November 10, 2025
- Large Document Support: Better handling of documents with many pages using smart chunking
- Validation Tracking: Track which values have been manually validated or edited
- Improved Accuracy: Enhanced extraction accuracy with updated default model
November 3, 2025
- Large Document Processing: Improved extraction quality for documents exceeding context limits
- Confidence Scoring: When merging extractions, the highest confidence value is preserved
October 2025
October 27, 2025
- Performance: General performance improvements across the platform
October 13, 2025
- CSV Export: Fixed field ordering in CSV exports to match workflow configuration
- Results Ordering: Extraction results now consistently respect field order
October 6, 2025
- Nested Field Reordering: New endpoint to reorder fields within object-type fields
- Verification URLs: Job results now include verification URLs for easy access
- Bug Fix: Fixed handling of fields with similar names in nested structures
September 30, 2025
- Improved Extraction Pipeline: Redesigned extraction workflow for better reliability