It’s been an exciting ride since our last update and we’ve been busy with new features.
Over the past year, Parseur has rolled out powerful enhancements across AI capabilities, OCR performance, export flexibility, and overall user experience. From custom AI instructions to multilingual support, our team has focused on improving speed, precision, and scale for your document parsing workflows.
AI Parsing Engine V3: Smarter, Longer, More Accurate
Our latest AI parsing engine update delivers:
- Improved long document parsing: Now handles complex multi-page documents with better memory and layout awareness.
- Enhanced instruction following: Custom AI instructions now yield even more accurate field-level extractions.
- Better layout understanding, especially for documents with inconsistent or messy formatting.
Google Login & Sign-Up Now Available
You can now sign up or log in to Parseur using your Google account. No more juggling passwords.
Already have an account with your Gmail address? You can now switch to Google login for quicker, secure access.
Smarter AI with Metadata & Parameters
Now you can make your AI instructions even more powerful:
Reference custom API parameters using
Custom:
prefix→ Example:
Custom:supplier_name
Access document metadata in the instructions with
metadata:
prefix→ Fields supported: Sender, SenderName, Recipient, Subject, PageCount, etc.
New Integration: Parseur Connector for n8n (Beta)
We now support sending and receiving documents via n8n, the open-source automation tool.
Head to the n8n tab under Export in your mailbox settings to start.
This integration is currently in beta while under review by the n8n team.
New Searchable PDF Link in Metadata
You can now access a cleaned-up, searchable version of your scanned PDFs directly from metadata.
- Rotated pages correctly
- OCR-enhanced text for better searchability
- If OCR was not applied, it defaults to the original PDF
Custom AI Instructions for Every Field
You can now give your AI parser custom instructions at the field level, ensuring smarter, more accurate extractions.
When to use AI field instructions:
- Add context to clarify ambiguous data
- Limit choices (e.g., pick from a predefined list)
- Summarize a section of the document
- Translate text into another language
Enhanced AI Engine for Long Documents
Parseur now supports documents up to 100 pages long, with improved parsing accuracy and faster processing speeds. This makes it ideal for handling contracts, technical reports, or multi-page invoices.
New! Image File Parsing Support
We now support extracting text from image formats, including:
- PNG, JPEG, TIFF, BMP, GIF
Whether you're working with scanned receipts, ID cards, or camera-captured documents, Parseur now seamlessly pulls data.
Enable this in Mailbox Settings → Processing → Enable Image File Processing.
Improved OCR Accuracy for Tables
Our latest OCR upgrades specifically address row merging issues, delivering better precision in multi-row table parsing, which is especially valuable for financial statements and order summaries.
Automatic Detection of Number and Date Formats
Parseur now automatically detects locale-based formatting, including:
- Date formats: MM/DD/YYYY vs DD/MM/YYYY
- Number separators: 1,234.56 vs 1.234,56
This makes global data processing smoother than ever.
Custom File Downloads & Field Ordering
You can now reorder fields before exporting to:
- Excel
- CSV
- Google Sheets
Perfect for streamlining your workflows and ensuring consistency in reports.
Parseur Is Now Multilingual
The Parseur app and website are now available in 10+ languages, including:
🇫🇷 French | 🇩🇪 German | 🇨🇳 Chinese | 🇯🇵 Japanese | 🇳🇱 Dutch | 🇮🇹 Italian | 🇵🇹 Portuguese | 🇪🇸 Spanish
Language preferences can be changed in User Settings.
Note: Translations are powered by AI and may have minor inaccuracies.
New Document Format Support: .mbox
You can now parse .mbox
files, making processing concatenated emails, support logs, and archived inbox exports easier.
More intelligent AI with Better Spatial Awareness
The latest AI engine features:
- Enhanced document structure recognition
- Stronger adherence to custom field instructions
- Greater resilience for scanned and complex layouts
More Intelligent Data Extraction
We’ve improved handling of complex global formats, including:
- Currencies: $1,234.56, ₹-7,50,721.05, (1.234,56€)
- Dates: Better parsing for Chinese, Japanese, and Korean date formats
Improved PDF Parsing Performance
Parseur now processes large PDFs faster and more efficiently, reducing latency and improving accuracy on high-volume document loads.
Seamless Server Migration
We completed a zero-downtime infrastructure migration, enhancing Parseur’s scalability and reliability without impacting our customers.
Improved Document View & Metadata Handling
We refined how documents are displayed and tagged with metadata, improving search, auditability, and traceability across your workflows.
Last updated on