What’s New

DocuShare 8.1 is our most AI-powered release to date, delivering significant advancements in intelligent document processing, enterprise search, workflow automation, security, and AI integration. Building on the foundation established in DocuShare 8.0, this release expands the capabilities of Content Assistant and introduces new tools to help organizations automate document-intensive processes, improve information discovery, and securely connect enterprise content with modern AI platforms.

Key innovations in this release include AI-powered document classification and filing, multi-document analysis, hybrid AI search, support for the Model Context Protocol (MCP), advanced invoice processing capabilities, and enhanced security controls. Together, these enhancements help organizations reduce manual effort, improve data quality, accelerate business processes, and unlock greater value from their enterprise content.

Model Context Protocol (MCP) Server

DocuShare 8.1 introduces support for the Model Context Protocol (MCP), a standardized framework that enables secure integration between DocuShare and MCP-compatible AI assistants, applications, and agent-based workflows. By adopting this emerging industry standard, organizations can connect enterprise content and business processes to AI platforms without requiring custom point-to-point integrations. The MCP Server provides a secure interface that allows AI clients to interact with DocuShare content, metadata, collections, and administrative functions while honoring existing DocuShare security controls and access permissions. AI agents can dynamically discover document classes, custom metadata structures, and available actions, allowing interactions to be grounded in the organization's actual DocuShare configuration.

Key Features

Support for MCP-compatible AI assistants and applications
More than 50 tools across authentication, search, document management, collections, favorites, and user administration functions
Dynamic discovery of document classes and custom metadata
Fine-grained administrative controls for capability exposure
Full ACL enforcement using authenticated user permissions
Built-in protection against prompt injection and cross-site scripting (XSS) attacks
Immediate activation of configuration changes without requiring service restarts

Benefits and Impact

Accelerates AI integration initiatives
Reduces the need for custom development and maintenance
Enables secure AI-powered automation scenarios
Extends the value of enterprise content within emerging AI ecosystems
Preserves existing governance, compliance, and security policies

Note: MCP capabilities are disabled by default and require administrator configuration and a Content Assistant license.

Personal Access Tokens (PAT)

To support modern integrations and automated workflows, DocuShare 8.1 introduces Personal Access Tokens (PATs). PATs provide a secure authentication method for API integrations, CI/CD processes, automation platforms, and MCP-enabled AI clients. Users can generate and manage Personal Access Tokens directly from their DocuShare profile, allowing integrations to authenticate securely without relying on browser-based login workflows. PATs work alongside the new MCP Server and support secure service-to-service communication across a variety of integration scenarios.

Key Features

Secure token-based authentication for integrations and automation
User-managed token generation and revocation
Support for MCP clients and API-based workflows
Exchange of PATs for short-lived JWT tokens to reduce credential exposure
Configurable token lifecycle controls
Multiple active tokens supported per user

Benefits and Impact

Simplifies integration with external systems and automation tools
Improves security for API-based workflows
Reduces reliance on traditional username and password authentication
Supports modern development and deployment practices
Provides greater control over integration credentials

Note: PAT generation is available to all users. Using PATs with MCP workflows requires that the MCP Server be enabled by an administrator.

AI-Powered Content Intelligence & Automation Enhancements

License Note: Features in this section require an additional Content Assistant license, configuration, and sufficient Content Assistant Normalized Page Packs. Contact your Xerox DocuShare authorized reseller for details.

Content Assistant: AI Auto-Filer

DocuShare 8.1 introduces AI Auto-Filer, an intelligent document intake solution that automates the end-to-end processing of incoming content. By combining AI Document Classification, AI Data Extraction, and AI Search indexing into a single workflow, AI Auto-Filer helps organizations streamline document intake while reducing manual processing and indexing efforts.

As documents enter the repository, AI Auto-Filer can automatically identify the document type, extract business-critical information, and prepare content for Content Assistant Search. Each stage of the workflow can be configured independently, allowing organizations to tailor processing based on document type, business requirements, and governance policies. AI Auto-Filer also incorporates human-in-the-loop review controls, ensuring that documents requiring additional validation can be reviewed before metadata is committed or workflows continue. This combination of automation and oversight helps organizations balance efficiency, accuracy, and compliance.

Key Features

End-to-end AI-powered document intake workflow
Automatic document classification with confidence scoring
AI-driven metadata extraction using document-specific extraction templates
Automatic indexing for Content Assistant Search
Configurable workflow stages and processing options
Human review workflows for low-confidence classifications and extractions
Flexible routing and review assignment capabilities
Support for OCR processing of scanned documents and images
Independent configuration of classification, extraction, and indexing stages

Benefits and Impact

Reduces manual document processing and filing
Improves metadata accuracy and consistency
Accelerates invoice, contract, and records management workflows
Supports straight-through processing for high-confidence documents
Improves searchability and discoverability of enterprise content
Increases operational efficiency through automated document intake
Provides governance and quality controls through configurable review processes

Extraction Review Experience

DocuShare 8.1 introduces an enhanced extraction review experience that provides reviewers with a side-by-side view of document content and AI-extracted metadata. Users can compare extracted values against the source content, review confidence indicators, view AI sourcing explanations, and make corrections before the information is committed to the document properties.

Extraction Template Enhancements

New enhancements to the extraction template provide greater flexibility and control over AI-driven data capture. Administrators can choose whether extraction processes overwrite existing values or populate only empty properties, thereby supporting multi-stage document-processing workflows and progressive metadata enrichment. Per-field confidence thresholds allow organizations to apply different validation requirements based on the importance of individual data elements, helping focus review efforts on the information most critical to downstream business processes.

Note: Features in this section require a Content Assistant license, appropriate configuration, and sufficient Content Assistant page-pack capacity.

Content Assistant: AI Document Classification

DocuShare 8.1 introduces AI Document Classification, enabling organizations to automatically categorize documents based on their content as they enter the repository. Rather than relying on manual filing or user-selected document types, Content Assistant can intelligently analyze documents and assign the most appropriate document class. Classification can be applied automatically through content rules, helping ensure consistent organization, improved metadata quality, and more efficient downstream processing.

Key Features

Automatic document classification based on content analysis
Confidence-based classification controls
Dry-run mode for testing and validation
Support for custom document classes and subtypes
Classification audit trail stored with each document
Configurable page limits to manage AI consumption
Custom confidence guidance for document classes

Benefits and Impact

Improves document organization and consistency
Reduces filing errors and manual effort
Creates a stronger foundation for automation workflows
Enhances metadata quality and search accuracy
Accelerates document intake processes
Supports governance and compliance initiatives

Content Assistant: AI Multi-Document Analysis (Investigate Dataset)

DocuShare 8.1 introduces AI Multi-Document Analysis through the new Investigate Dataset capability. This feature allows users to analyze multiple documents simultaneously and ask questions across an entire set of selected content. Unlike traditional search, which focuses on locating information within individual documents, Investigate Dataset enables Content Assistant to analyze complete documents collectively, helping users identify patterns, inconsistencies, trends, risks, and relationships across large collections of content.

Key Features

Cross-document analysis and comparison
Natural language questioning across multiple documents
Full-document context for deeper analysis
Reusable Content Assistant instructions and prompts
Dynamic token optimization to reduce consumption
Session export to structured PDF reports
Executive summary generation

Benefits and Impact

Accelerates research and investigative work
Simplifies contract and compliance reviews
Helps identify inconsistencies across related documents
Improves decision-making through deeper content analysis
Reduces the time required for manual document comparison
Enables new AI-driven knowledge discovery use cases

Content Assistant: Hybrid AI Search

DocuShare 8.1 introduces a major enhancement to AI Search by combining semantic search with structured metadata filtering. Rather than relying solely on vector search, Content Assistant can intelligently determine when metadata, document properties, identifiers, dates, or other structured information should be used to narrow the search scope before performing semantic analysis.

This hybrid approach improves accuracy, scalability, and transparency while maintaining the conversational search experience users expect from Content Assistant.

Key Features

Combines semantic vector search with metadata-driven filtering
Automatically generates Solr or IDOL pre-filters when appropriate
Supports identifiers, document numbers, customer IDs, dates, and names
Visual indicators identify when pre-filtering was applied
Filter Details view displays generated filters in plain language or native syntax
Intelligent wildcard handling for partial matches
Automatic fallback to semantic search when filters return insufficient results

Benefits and Impact

Improves search accuracy for business-critical content
Delivers better performance in large repositories
Helps users understand how results were generated
Reduces time spent locating specific information
Scales more effectively across repositories containing millions of documents
Combines the strengths of traditional search and AI-powered discovery

Content Assistant: Premium AI Model Selection

DocuShare 8.1 introduces Premium AI Model Selection, providing organizations with greater flexibility when balancing AI performance, reasoning capabilities, and consumption costs. Authorized users can select from multiple AI model tiers when performing Content Assistant operations, allowing organizations to apply more advanced reasoning models to complex content while using standard models for everyday processing.

Key Features

Multiple AI model tiers available
Premium model selection for Content Assistant workflows
Support across search, summarization, extraction, and AI Auto-Filer
Visibility into credit consumption multipliers
Per-user and per-group access controls

Benefits and Impact

Improves accuracy for complex documents and use cases
Provides greater flexibility in AI deployment strategies
Enables cost optimization based on business requirements
Supports advanced reasoning for contracts, technical content, and specialized documents
Gives administrators greater control over AI resource consumption

Note: Premium model access requires administrator configuration and appropriate licensing.

Content Assistant: Search and Chat Advances

DocuShare 8.1 introduces significant enhancements to the Content Assistant search experience, making it easier for users to find information, continue research, and understand how AI-generated responses are produced. New conversational capabilities provide greater flexibility, while enhanced transparency helps users evaluate the quality and completeness of results.

Key Features

Toggle between traditional keyword search and AI-powered Content Assistant Search directly from the search bar
Save and resume Content Assistant conversations
Export sessions as PDF documents for sharing, collaboration, and audit purposes
Revert conversations to earlier points and explore alternate lines of inquiry
Enhanced answer transparency, including source counts, question restatement, and page consumption tracking
Improved relevance detection that alerts users when available information may be insufficient to fully answer a question
Administrative controls that determine whether Content Assistant responses are generated exclusively from DocuShare content or may leverage general AI knowledge

Benefits and Impact

Helps users find information faster using natural language queries
Improves trust and confidence in AI-generated responses
Enables teams to preserve and share research sessions
Provides greater visibility into how answers are generated
Supports more informed decision-making by identifying potentially incomplete results

Content Assistant: AI Summarization in All 11 Supported Languages

DocuShare 8.1 expands AI Summarization language support to all 11 DocuShare-supported languages, making AI-powered content summarization more accessible for global organizations.

Key Features

Support for all DocuShare-supported languages
Language selection for AI-generated summaries
Available through both user-initiated and content rule-driven summarization

Benefits and Impact

Supports multinational and multilingual organizations
Improves accessibility of content across regions
Enables consistent AI-powered experiences regardless of language
Helps users quickly understand content in their preferred language

Content Assistant: OCR Support for Classification

DocuShare 8.1 extends AI Document Classification to scanned documents and image-based content through integrated OCR processing. When OCR mode is enabled, Content Assistant can extract text from scanned PDFs and images before performing classification.

Key Features

OCR support for scanned PDFs and image files
Classification of documents without an existing text layer
Flexible OCR processing options
Integration with existing classification workflows

Benefits and Impact

Expands automation to paper-based and scanned content
Improves classification accuracy for image-based documents
Reduces manual handling of scanned records
Delivers a more consistent document intake experience across content types

Note: Requires a Content Assistant license and sufficient Content Assistant Normalized Page Packs.

Invoice Processing & Metadata Enhancements

DocuShare 8.1 introduces a comprehensive set of metadata and data management enhancements designed to support invoice processing, contract management, records management, and other document-centric business processes. These capabilities enable organizations to manage structured business data directly within DocuShare while improving automation, validation, and workflow efficiency.

New capabilities include support for multi-value table data, intelligent field population, reference data lookups, confidence-based extraction review, and automated synchronization of external business data. Together, these enhancements help reduce manual data entry, improve accuracy, and provide greater visibility into business-critical information.

Multi-Value Fields (ItemsTable Data Type)

DocuShare 8.1 introduces the ItemsTable property type, allowing organizations to store structured, tabular data directly within document metadata. This new property type supports multiple rows and columns of information, making it ideal for invoice line items, contract participants, approval histories, parts lists, and other repeating business data. Rather than relying on external spreadsheets or custom applications, organizations can now manage structured business data alongside the documents that support it.

Key Features

Multi-row, multi-column table data stored directly within document properties
Administrator-defined table structures
Support for string, text, integer, and decimal column types
Required field validation
Search, sorting, and pagination capabilities
Inline editing and row management
Multi-select property editing support
Workflow and routing task integration
Content rule support for data presence conditions

Benefits and Impact

Simplifies invoice and contract processing
Reduces reliance on external spreadsheets
Keeps business data connected to supporting documents
Improves visibility into structured information
Supports more advanced workflow and approval processes
Enhances document-centric business applications

Common Use Cases

Invoice line item management
Purchase order validation
Contract participant tracking
Approval and signature chains
Product and parts management
Asset and inventory documentation

AutoFill Related Fields

DocuShare 8.1 introduces AutoFill, an enhancement that automatically populates related document properties when users select values from AutoComplete-enabled fields. For example, selecting a vendor name can automatically populate associated information such as addresses, phone numbers, account numbers, or payment terms. AutoFill can also support multi-level population chains, reducing repetitive data entry and improving consistency.

Key Features

Automatic population of related metadata fields
Integration with AutoComplete lookup tables
Cascading field population across multiple data sources
Automatic clearing of dependent values when source values change
Flexible administrator configuration

Benefits and Impact

Reduces manual data entry
Improves metadata consistency and accuracy
Accelerates document indexing and review
Reduces user errors
Simplifies business process workflows

AutoComplete Lookup Table View

DocuShare 8.1 enhances AutoComplete fields with a new Lookup Table View, providing users with a searchable interface for browsing available records. Users can review available values, search records, sort results, and select entries directly from a dedicated lookup window. Associated AutoFill values are populated automatically once a selection is made.

Key Features

Searchable lookup table interface
Sorting and pagination support
Browse and select available records
Integration with AutoComplete and AutoFill
Visual indication of currently selected records

Benefits and Impact

Improves usability for large datasets
Helps users locate the correct values quickly
Reduces data entry errors
Enhances overall user experience

Table Reference Data Type

The new Reference property type provides direct visibility into related business data stored within custom database tables. Reference fields display linked information in a read-only format, allowing users to review associated records without leaving the document. Reference fields can also dynamically filter displayed information based on values already associated with the document, helping present only the most relevant records.

Key Features

Direct linkage to custom database tables
Searchable and paginated reference views
Dynamic filtering based on document metadata
Read-only access to related business information
No manual data entry required

Benefits and Impact

Provides better context for business decisions
Improves visibility into related records
Supports invoice and purchase order validation processes
Reduces the need to access external systems
Simplifies information review and verification

Confidence Scoring and Human-in-the-Loop Review

DocuShare 8.1 expands AI Data Extraction with confidence scoring and enhanced review workflows. Individual extracted values can now be evaluated against administrator-defined confidence thresholds, allowing organizations to determine when human review is required. This approach helps balance automation efficiency with accuracy requirements, particularly for business-critical content such as invoices, contracts, financial records, and regulated documents.

Key Features

Confidence scoring for extracted values
Per-field confidence thresholds
Automated routing for review and validation
Streamlined review interface
Support for exception-based processing

Benefits and Impact

Improves trust in AI-generated metadata
Reduces manual review effort
Focuses attention on higher-risk extractions
Improves overall data quality
Supports compliance and governance requirements

Flat File Import and Automated CSV Ingestion

DocuShare 8.1 introduces new capabilities for importing and synchronizing business data into custom database tables. Organizations can now load data from external systems, including ERP platforms, vendor databases, cost center lists, and product catalogs. Reusable transformation profiles simplify the mapping process, while automated ingestion services can monitor directories and process incoming data files automatically.

Key Features

Import data directly into custom database tables
Reusable transformation profiles
Automatic field mapping assistance
Support for append, update-or-append, and replace-all load strategies
Automated ingestion through monitored directories
Processing notifications and reporting
Optional results files for audit and troubleshooting

Benefits and Impact

Keeps reference data synchronized with external systems
Reduces manual maintenance of lookup tables
Improves accuracy of AutoComplete and AutoFill data
Supports large-scale business process automation
Simplifies ongoing data management and administration

Transformation Profiles

Transformation Profiles provide a reusable framework for mapping incoming CSV files to custom database tables. Administrators can define mappings once and reuse them across multiple import processes, reducing setup effort and improving consistency.

Automated Ingestion

Automated ingestion capabilities allow DocuShare to monitor designated locations for incoming files and process them according to predefined transformation rules. This helps organizations maintain up-to-date reference data while minimizing manual intervention.

Note: CSV is the supported format for automated ingestion. JSON is supported when included in CSV fields to populate ItemsTable properties.

Content Rule Enhancements

DocuShare 8.1 introduces several enhancements to Content Rules, making automation easier to configure, manage, and maintain. New organizational tools, improved usability, expanded AI capabilities, and support for structured data workflows help administrators build more powerful business processes while reducing complexity. These enhancements improve the overall administration experience and provide greater flexibility when designing document-centric workflows.

Tags Management for Content Rules

DocuShare 8.1 introduces support for tags in Content Rules, enabling administrators to organize and categorize automation rules by business function, department, document type, workflow, or other organizational requirements. Tags provide a simple way to manage large rule libraries and make it easier to locate, maintain, and audit automation configurations.

Key Features

Create and assign tags to Content Rules
Filter rule lists by tag
Apply tags individually or in bulk
Centralized tag management interface
Visibility into tag usage across the system
Foundation for future tagging support across additional object types

Benefits and Impact

Simplifies management of large rule libraries
Improves administrator productivity
Makes automation easier to organize and maintain
Supports governance and documentation efforts
Reduces time spent locating and updating rules

Redesigned Content Rule Action Selector

DocuShare 8.1 delivers a completely redesigned Content Rule Action Selector that makes it easier to discover and configure automation actions. The previous flat list of actions has been replaced with a categorized, searchable interface that helps administrators quickly locate the actions they need and provides additional context for each option.

Key Features

Categorized action organization
Searchable action selection interface
Real-time keyword filtering
Dynamic descriptions for selected actions
Logical grouping by function
Automatic support for custom actions

Benefits and Impact

Simplifies rule creation and configuration
Reduces administration time
Improves discoverability of available actions
Lowers the learning curve for new administrators
Makes complex automation easier to build and maintain

ItemsTable Property Conditions

With the introduction of the ItemsTable property type, Content Rules can now evaluate whether structured table data exists within a document and trigger actions accordingly. This capability allows organizations to build workflows that respond to the presence or absence of structured business data such as invoice line items, approval records, or contract-related information.

Key Features

Support for ItemsTable properties within Content Rule conditions
Conditions based on whether table data exists or is empty
Visual validation of conflicting rule conditions
Integration with workflow and routing scenarios

Benefits and Impact

Enables more intelligent workflow automation
Supports invoice and financial processing scenarios
Expands automation options for structured business data
Improves workflow flexibility and control

Content Assistant OCR Actions

DocuShare 8.1 expands automation capabilities by introducing Content Assistant OCR actions directly into the Content Rule framework. Organizations can now automatically process scanned documents and image-based content as part of broader workflow automation initiatives. In addition, several Content Assistant actions now support OCR fallback processing, ensuring documents without extractable text can still participate in AI-powered workflows.

Key Features

Dedicated Content Assistant OCR action
Automatic creation of searchable PDF renditions
Support for PDF and common image formats
OCR fallback support for AI-powered actions
Integration with AI Data Extraction
Integration with AI Auto-Filer
Integration with AI Search indexing
Support for premium AI model selection where applicable

Benefits and Impact

Extends automation to scanned and image-based content
Reduces manual OCR processing requirements
Improves accessibility and searchability of scanned documents
Expands the effectiveness of AI-powered workflows
Creates more consistent document processing experiences across content types

OCR Fallback Processing

OCR fallback allows Content Assistant workflows to automatically perform optical character recognition when a document contains no extractable text. This helps ensure that scanned documents can still be classified, indexed, summarized, and processed by AI-powered workflows without requiring separate OCR steps.

Note: Content Assistant OCR capabilities require a Content Assistant license and sufficient Content Assistant Normalized Page Packs.

Authentication and Security

DocuShare 8.1 introduces a range of security and authentication enhancements designed to help organizations strengthen access controls, protect sensitive information, and support modern identity management strategies. These improvements provide additional layers of protection while maintaining the flexibility required to support diverse deployment environments and integration scenarios. New capabilities include multi-factor authentication for privileged accounts, expanded SAML authentication options, enhanced upload security, encryption improvements for AI-generated data, and additional controls for protecting content accessed through the Office Connector.

Multi-Factor Authentication (MFA) for Privileged Accounts

DocuShare 8.1 introduces built-in multi-factor authentication (MFA) for privileged administrative accounts, providing an additional layer of protection for users responsible for managing the platform.

When enabled, authorized administrators must complete an email-based one-time password (OTP) verification step during authentication. MFA enrollment and recovery options are designed to provide enhanced security while minimizing disruption to administrative access.

Key Features

Email-based one-time password (OTP) authentication
MFA support for Site Admin, Content Admin, and Account Admin roles
Recovery code generation during enrollment
Configurable expiration periods and retry limits
Device trust options to reduce repeated authentication prompts
Administrative policy controls for MFA management

Benefits and Impact

Strengthens protection for privileged accounts
Reduces the risk of unauthorized administrative access
Supports security and compliance initiatives
Provides additional safeguards for sensitive system functions
Helps organizations implement stronger authentication practices

Note: MFA is not applied to SAML-enabled sites where authentication is managed by the organization's identity provider.

SAML + DocuShare Dual Authentication

DocuShare 8.1 expands authentication flexibility by allowing selected accounts to use traditional DocuShare authentication even when SAML single sign-on is enabled across the site. This capability is particularly valuable for service accounts, integrations, and automated processes that cannot participate in browser-based SAML authentication workflows.

Key Features

Support for mixed authentication environments
Per-user authentication configuration
Service account compatibility
Logging of authentication bypass activity
Support for integrations and API-based tools

Benefits and Impact

Simplifies integration with external applications
Supports automation and service account scenarios
Reduces deployment complexity for organizations using SAML
Maintains flexibility without sacrificing governance

Common Use Cases

API integrations
Import and export utilities
Automated business processes
Third-party integration tools
Service account authentication

Guest Access on SAML-Enabled Sites

DocuShare 8.1 enhances support for public-facing content by allowing guest access to coexist with SAML authentication. Organizations can provide anonymous access to designated content while requiring authentication only when users attempt actions that require additional permissions. This flexibility allows organizations to balance security requirements with public information-sharing needs.

Key Features

Anonymous access support on SAML-enabled sites
Authentication only when required
Multiple authentication deployment models
Flexible access control options
Support for public content repositories

Benefits and Impact

Improves user experience for public-facing sites
Simplifies access to publicly available content
Reduces unnecessary authentication prompts
Supports a wider range of deployment scenarios
Provides greater flexibility for content sharing initiatives

Note: Guest access availability depends on site licensing.

File Content Validation

DocuShare 8.1 introduces enhanced upload security through file content validation. Uploaded files are inspected to verify that their actual content matches the file type being submitted. This protection helps prevent malicious or improperly renamed files from entering the repository, reducing potential security risks across all supported upload channels.

Key Features

Validation of file content against file extensions
Protection against disguised or renamed files
Support for common document, image, archive, and media formats
Coverage across all major upload methods
Configurable validation controls

Benefits and Impact

Strengthens repository security
Reduces exposure to malicious file uploads
Helps protect users and business content
Improves confidence in uploaded content
Supports security best practices

Encryption at Rest for AI Vector Data

Organizations using Content Assistant and document encryption can now extend encryption-at-rest protections to vector data stored for AI Search operations. This enhancement applies the same encryption approach used for protected content to AI search data, helping organizations maintain consistent security practices across both traditional content repositories and AI-powered services.

Key Features

Encryption-at-rest for AI vector data
Integration with existing DocuShare encryption capabilities
Protection of Content Assistant search data
Automatic encryption for newly indexed content
Support for migration of existing vector data

Benefits and Impact

Strengthens protection of AI-related content data
Supports security and compliance requirements
Extends encryption coverage across AI-powered services
Helps organizations maintain consistent data protection policies

Note: Content encryption is a separately licensed feature. Encryption is enabled by default for DocuShare Online customers.

Office Connector Download and Print Restrictions

DocuShare 8.1 introduces new controls that allow administrators to restrict downloading and printing of content viewed through the Office Connector and OnlyOffice viewer. These controls help organizations protect sensitive information by limiting how read-only users can interact with documents while maintaining access to the content itself.

Key Features

Restrict document downloads for read-only users
Restrict document printing for read-only users
Enforcement within the OnlyOffice viewer
Coverage of browser and API-based export actions
Administrator-controlled configuration

Benefits and Impact

Improves protection of sensitive information
Supports document security policies
Reduces the risk of unauthorized content distribution
Enhances governance and compliance controls
Provides greater administrative control over content access

Note: These restrictions apply to content viewed through the OnlyOffice viewer and Office Connector experience.

Removed Features

The following features and platform support have been removed in DocuShare 8.1:

WebEx / Online Meetings integration
LiquidOffice forms integration
SharePoint Export Agent (SPExporter)
WebDAV support

Upgrade Paths

DS 7.5 + DS 7.5p2 → DS 7.7p1 → DS 8.1
DS 7.6 + DS 7.6p6 → DS 8.1
DS 7.7p1 → DS 8.1
DS 8.0 → DS 8.1

System Requirements

Operating Systems

Windows Server 2019, 2022, 2025 (all 64-bit)
Ubuntu 22.04
Red Hat Enterprise Linux 9 (64-bit)

Application Server

Tomcat 9.0.120

Java Runtime

Amazon Corretto JDK 21

Web Server

Apache 2.4.6
IIS 10

Databases

Microsoft SQL Server 2022, 2019
Microsoft SQL Server 2022 Express Edition
Microsoft Azure SQL Database (New)
PostgreSQL 13, 15, 17, 18
Oracle 19c, 23ai

Search Engine

Solr 9.10.1

PDF Library

OpenPDF 3.0.1 (replaces iText 5.0.1)

OCR

Updated OCR component with digital signature preservation

Web Browser Support

Google Chrome (latest)
Microsoft Edge–Chromium (latest)
Safari (latest)

Licensing

A new license is required for DocuShare 8.1.
DocuShare supplies a 30-day grace period to apply the new license after the upgrade.

Installation Notes

pgVector Add-on Installer

DocuShare 8.1 introduces a dedicated add-on installer for pgVector, replacing the previous manual database setup process. Supports existing PostgreSQL installations (on the same or a remote server) and new PostgreSQL installations. Supports PostgreSQL versions 13 through 18. Vector database connection passwords are automatically encrypted during installation.

Solr

Multiple DocuShare instances on a single VM with Solr are not supported.
The Solr engine starts automatically after installation on Linux and Windows. If a Linux server is restarted, Solr must be started manually before starting DocuShare.
Only a net new install provides the Solr engine. Upgrading from a version running the IDOL Service retains the IDOL engine.
Solr reindex is required after upgrading to 8.1.

Note: If you need assistance migrating from IDOL to Solr (KeyView/IDOL to Tika/Solr), contact Xerox customer support.

pgVector

The pgVector add-on for Content Assistant semantic search is installed separately after the main DocuShare installation.
PostgreSQL is required for pgVector support.

Additional Notes

Xerox recommends that all customers running the IDOL search engine plan their migration to Solr. IDOL support is being phased out and will reach end-of-life in a future release. If you need assistance migrating from IDOL to Solr (KeyView/IDOL to Tika/Solr), contact Xerox DocuShare customer support or your Xerox representative.

XSLT support is scheduled for removal in November 2026. Customers using XSLT-based customizations should begin planning their migration now.

Bug Fixes in This Release

AR Number	SPAR Number	Description
46688	1626	Scanning from a Xerox ConnectKey device could fail or hang when Auto-Login was enabled, causing the scan job to time out instead of filing the document.
58893	4735	Certain wildcard searches (for example, patterns containing a hyphen) returned no results when searching across all fields with the Solr search engine.
60571	5324	OCR processing could remove or blank out the pages of PDF documents containing digital signatures.
63169	5751	Administrators who belonged to a group nested within the Site Administrators group received a "Permission denied" error when attempting to view or download log files.
63245	5766	Entering an emoji or other extended Unicode character in an approval task comment could fail when written to a PDF form and halt the rest of the workflow.
63643		AI Search (Content Assistant) could return no results for collection-scoped searches, and some non-administrator users could not retrieve any AI Search responses; result handling and relevance for both search engines were also improved.
64581	5990	AI Search (Content Assistant) responses could display raw formatting markup, such as LaTeX notation, instead of clean, readable text.
64612		The document abstract (generated summary) property was not populated after a document was indexed.
65217	6192	In the embedded document viewer and editor, the editing area did not account for the page header and footer, requiring users to manage two separate scroll bars to view an entire document.
65238	6165	Scanning from a Xerox ConnectKey device failed when SAML authentication was enabled on the DocuShare server. Users could browse DocuShare collections from the device, but the scan job was canceled and the device reported a lack of system resources.
65354	6078	Recipients of a routing slip that allowed changes were unable to upload a new version of a reviewed document, even though the routing slip granted them permission to do so.
65655	6251	The Related Documents list could return extra documents that did not match the relationship criteria when the matched value contained special characters such as periods.
66388	6487	The results panel for an AI-assisted (semantic) search could become excessively wide, introducing an oversized horizontal scrollbar when results contained long code blocks, tables, or unbroken text.
66516	6378	An advanced search limited to a specific custom object type could also return documents of other object types when one object type's display label matched another object type's internal name.
66812	6386, 6429, 6661, 7632	Addressed security vulnerabilities (CVE-2025-54988 and CVE-2025-66516) in the third-party Apache Tika text-extraction component by upgrading the search subsystem and isolating text extraction in a hardened external Tika service.
67320	7843	The command-line administrator password reset could fail with a database driver error and was unable to reset the administrator password.
70317	6705, 6181, 6319	If the conversion service's temporary working directory was removed, it was not recreated automatically, which could flood the conversion log with errors and cause high CPU usage and significantly degraded site performance.
70532	6895	Searching on a single-character menu property value could return more documents in the results than expected.
72220	7219	The multi-select bulk download could bypass the configured download size limit, allowing an excessive number of files to be selected and potentially exhausting server memory.
74605	7410	The Track Changes option was not available in the OnlyOffice editor when editing a document.
75378	7422	Newly indexed documents could intermittently be found in search by administrators but not by non-administrator users until a reindex was performed.
82184	7666	In the Xerox ConnectKey for DocuShare scan application, the "Scan another document" prompt shown after a scan appeared as plain text instead of a clickable link, preventing users from scanning additional documents.
83031	8024	Security scans flagged outdated Apache Commons FileUpload libraries (versions 1.4 and 1.5) that remained in the installation after upgrading, reporting them as vulnerable to a denial-of-service issue (CVE-2025-48976). All bundled copies of Apache Commons FileUpload are now version 1.6, and the superseded 1.4 and 1.5 files are no longer installed.
83497		AI Data Extraction does not support certain property datatypes in this release. Table/line-item (ItemsTable) properties, and other non-scalar types (object Reference, XML, LDAP), are not extracted correctly and should not be selected as data-extraction template fields; Password properties are never valid extraction targets. Full support for ItemsTable line-item extraction is planned for a future release.
84156	8305	Enabling Firewall Mode prevented DocuShare from starting. Startup failed with a NullPointerException reported for DSWebServlet, so sites that required Firewall Mode — for example, to use DocuShare Federation — were unable to bring the service up. DocuShare now starts correctly when Firewall Mode is enabled.

Known Issues

AR Number	Description	Work Around
13266	When running encryptContentStore (.bat/.sh) with -encrypt -logfile <path>, the command output is not written to the full file path given by -logfile. The encryption operation completes normally; only redirection of the log output to the custom path is affected.	To capture the command output to a log file, edit config\log4j2.xml and add the dscommand-log appender so output is written to the existing logs\dscommand.log. Change <appender-ref ref="console-log"/> to <appender-ref ref="console-log"/><appender-ref ref="dscommand-log"/>.
35909	When joining a federation, you may experience the following error: Sorry. Your request could not be processed. 1136: This federation member has failed to respond and will not be touched for a configured amount of time. If that occurs, it is because of a port blockage for connection. You will need to create inbound and/or outbound rules in the firewall that will allow the port to connect the two servers.	This error occurs when the port used for federation communication is blocked by a firewall. Create inbound and/or outbound firewall rules on both servers to allow traffic on the federation connection port, then retry joining the federation.
36234	When creating a content rule, you are allowed to create actions that without filling in all the mandatory fields, because the current UI framework cannot handle mandatory validation and still allow the back button to work. This change is significant and current direction to the user is - if you don't fill in the fields correctly - regardless of mandatory status - the action will not work. So the user should take care to fill everything out correctly.
50588	On some systems the Solr search service intermittently fails to restart automatically after the scheduled daily backup routine completes.	Use the provided helper script start_solr.ps1: (1) place start_solr.ps1 in the <DSHome>\bin folder; (2) in your daily backup batch file, replace 'net start docushare-solr' with 'PowerShell.exe -ExecutionPolicy Bypass -File "<DSHome>\bin\start_solr.ps1"'. The script starts the Solr service, verifies it is healthy, and if not, stops and restarts it (up to 5 attempts), waiting progressively longer each retry to allow Solr enough time to start.
55747	** Impacts Office Connector Only ** PDF Editing is a new feature of the service used for the Office Connector. When "strict" co-editing mode is enabled, the collaboration tab is not shown as expected. This has been reported to the application team. As soon as a fix is available, our instance will be upgraded, and this issue will be resolved for all customers.
56889	Not all strings are currently translated within the Office Connector. This will be corrected in a future release.
58018	The Content Rule creation wizard uses POST-based navigation between steps. The browser back button cannot replay a POST submission from cache, so it returns ERR_CACHE_MISS. This is standard browser behavior for POST-based forms; controlling it from server side would require either disabling caching (security/performance impact) or refactoring the wizard to a Post-Redirect-Get pattern (significant rework not justified by the symptom). The wizard provides a Previous button for in-flow navigation, which is the supported and documented way to move backward through steps. Users hitting the browser Back button instead are bypassing that supported affordance. Status: Working as Intended.	Use the in-wizard Previous button to navigate back through Content Rule wizard steps. Do not use the browser's Back button.
58879	When attempting to index documents greater than or equal to 500MB using the Content Assistant fails with the default settings. The log files contain the following error: ERROR - Unexpected error in worker thread, continuing: Java heap space	Increasing the memory allocation from -Xmx512m to -Xmx3072m in the monitor.xml file will resolve the issue.
59585	When you provide data-extraction instructions (a prompt) to the Content Assistant in a PFIGS language (Portuguese, French, Italian, German, Spanish), the extracted results are generally returned in that same language. The behavior is not fully consistent — results are sometimes returned in English — because the output language follows the language of the prompt rather than an explicit setting.	To control the language of the extracted results, state the desired output language explicitly in your extraction instructions — e.g., add "Return the extracted text in English." (or the language you require) to the prompt.
59596	When SAML single sign-on is enabled and the session cookie's SameSite attribute is set to Strict, you cannot reach the DocuShare home page after authenticating. SAML relies on a cross-site redirect from the identity provider, and a Strict SameSite cookie is not sent on that redirect, so the session is not recognized.	Do not use SameSite=Strict when SAML is enabled. Use Lax (the default), which allows the session cookie on the identity-provider redirect while still providing cross-site protection.
59774	You are unable to remove the index on a document attached to a discussion topic when using the Remove Content Assistant Index using "Selected Collections and All Levels of Content" option. To remove the index on a document attached to a discussion topic you must go to the document itself and use the "Remove Content Assistant Index" option.	To remove the Content Assistant index from a document attached to a Discussion topic, open the document itself and use its "Remove Content Assistant Index" option, rather than the collection-level "Selected Collections and All Levels of Content" option.
59842	Disabling Content Assistant does not prevent / hide the Content Assistant Content Rule actions. This allows the content rules to be available for troubleshooting even if the Content Assistant was disabled. Note that Other UI actions are hidden when Content Assistant is disabled.
59960	In order to improve the usage of Routing Tasks for the majority case, some action buttons were purposefully left off. This is specifically the case for the "Do Later" buttons. You can simply use the Next button in the page navigation to move to the next item if you wish to review that task "later". For other functionality that you might need for your workflow, you can use the Document > Properties > Routing menu option to view a more complete view.
60033	In order to improve the usage of Routing Tasks for the majority case, some action buttons were purposefully left off. This is specifically the case for the "Do Later" buttons. You can simply use the Next button in the page navigation to move to the next item if you wish to review that task "later". For other functionality that you might need for your workflow, you can use the Document > Properties > Routing menu option to view a more complete view.
61143	When configuring the subscription using dsemailservice.bat/.sh with smtp.sendgrid.net, the Subscription type is set to SMTP-TLS (as expected) in the config output. However, after starting DS and checking the Subscription page in the DS Admin UI, the Subscription type appears as SMTP only. The setting is set correctly. This is a display only issue that will be fixed in a subsequent release,
61690	The search service health-check command described in the command-line documentation does not return the expected collection status in this release. The documentation will be updated to reflect the current command behavior.	The solr healthcheck command does not correctly report the DocuShare collection status in this release. To check search-collection health, query the Solr Collections API directly (no credentials or HTTPS are required in the default configuration): curl "http://localhost:8983/solr/admin/collections?action=CLUSTERSTATUS&collection=DocuShare&wt=json". This returns the full collection status — shard/replica state, leader, and node information — equivalent to what the health-check command was intended to display.
63285	After upgrading DocuShare (e.g., to 8.0) on a server running IDOL with multiple content stores, DocuShare may fail to restart. This occurs when the license key file still contains the license key from the previous DocuShare version.	Update the license key file with the key for your upgraded DocuShare version: open the license properties file in the DocuShare configuration directory, replace the previous version's license key with your current version's key, and restart DocuShare.
64741	On some systems the default timeout is not long enough for the services to come up. The timeout setting should be increased in those cases. See Work Around.	Increase the service startup timeout in the Windows registry: (1) open Registry Editor and go to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control; (2) set the ServicesPipeTimeout DWORD value to 120000 (ms); (3) restart the server and confirm the DocuShare service starts.
65216	When editing a fillable PDF form in the integrated document viewer, changes to form fields are not always captured by the AutoSave function. In some cases this happens because the form is being edited in preview mode rather than edit mode. Less commonly, a specific PDF may have characteristics that prevent the third-party viewer component from saving edits.	First, make sure the document is open in edit mode rather than preview mode before entering data. If your entries are still not saved automatically, use the Save action, or switch back to view mode to force a save. If the values still do not save, the PDF itself may be preventing editing in the viewer; we are working with the component vendor to identify the cause. In the meantime, please report any affected documents to Support.
72240	If you configure the quicksearch right-hand panel such that the list of items that would be displayed is large, the site can experience out-of-memory errors and crash.	You should configure the quicksearch right-hand panel such that it will display only a manageable number of items. The panel is not usable when there are a large number of items in it anyway.
80510	When an administrator creates a new Federation, the operation appears to succeed but the configuration is not retained. After reloading the page or returning to the settings, the newly created Federation is no longer present.	Restart the DocuShare server, then create the Federation again. After the restart, the Federation configuration is retained as expected.
82381	AI Search (Content Assistant search) is not supported for documents stored on an Archive Server; archived documents are not returned in AI Search results.
82518	When a document contains a mix of English (or other Western) text and Japanese or other East Asian (CJK) characters, a full-text content search for the East Asian terms may return no results, even though those terms appear in the document. Searches for the Western-language portions of the same document are not affected.
82592	On a Windows server where DocuShare's web server (Tomcat) is fronted by the IIS web-server bridge, applying the recommended IPv4-preference setting (-Djava.net.preferIPv4Stack=true) to the Tomcat service can make DocuShare unreachable from a browser. Users see a "Service Temporarily Unavailable" page even though the server is running. This happens because the internal connector between IIS and Tomcat is configured to listen on an IPv6 address, which an IPv4-only Java process cannot use. DocuShare remains reachable directly on the Tomcat port (for example http://localhost:8080/...), but not through the normal web-server URL.	Pin both ends of the IIS-to-Tomcat (AJP) connection to the IPv4 loopback address: 1. Tomcat connector — edit <DSHome>\tomcat\conf\server.xml, find the AJP 1.3 Connector, and change the bind address from :: to 127.0.0.1: <Connector port="8009" protocol="AJP/1.3" address="127.0.0.1" secretRequired="true" secret="..." redirectPort="8443"/> (Leave the existing secretRequired/secret values unchanged.) 2. IIS bridge worker — edit <DSHome>\tomcat\conf\ntiis\workers.properties and set the worker host explicitly to IPv4 loopback: worker.ajp13.host=127.0.0.1 3. Stop and start DocuShare, then retry the web-server URL. This aligns the connector and the bridge with the IPv4-only Tomcat JVM. It also improves security, since the AJP port is no longer exposed on all network interfaces. The IPv4-preference setting (-Djava.net.preferIPv4Stack=true) can remain in place on all services.
82779	The InstallShield (ISMP) wizard framework chrome -- the Back/Next/Cancel navigation buttons and the title bar -- is rendered by the installer's bundled JRE, whose fontconfig has no CJK/Thai font mapping, so those strings render as tofu boxes (□). The DocuShare-authored panels (Welcome / EULA / Release Notes) render correctly via LocaleFontHelper (+ the welcome-panel putClass fix and the BOM-strip fix). This framework surface is not reachable from DocuShare application code; resolving it requires a fontconfig.properties / bundled-JRE font change in the installer JRE.	None required to complete installation. The navigation buttons stay in their standard fixed positions (Back at lower-left; Next / Cancel at lower-right) and are fully functional -- only the button labels are affected, and the page content is localized. Selecting English for the installer avoids the cosmetic issue entirely.
83148	When a document is converted to Markdown by the AI content tools (the get_document_markdown capability), the conversion is a text-extraction process and does not perform full page-layout analysis. As a result:PDF files: Tables and multi-column layouts are flattened into sequential text and are not reproduced as Markdown tables. Label/value pairs may lose their pairing. Word (DOCX) files with embedded images/screenshots: Text extracted from the embedded images can appear as scrambled content interleaved with the body text, embedded image file names can be rendered as document headings, and repeated page header/footer text may appear more than once. The extracted textual content of the document remains available; the limitation affects the structural fidelity of tables and the handling of content embedded in images.
83176	When the Content Encryption license is enabled on a DocuShare system that is already running with Content Assistant (AI indexing) configured, the new encryption setting is not applied to documents indexed afterward. Content stored in the vector database continues to be saved unencrypted (as plaintext) even though the Content Encryption license shows as enabled on the Site Configuration page. This occurs because the encryption state is read once when the AI indexing service starts and is not refreshed when the license is changed while the system is running.	Restart the DocuShare service after enabling the Content Encryption license. Once DocuShare has restarted, documents indexed thereafter are encrypted correctly — their content is stored as AES-128 encrypted, Base64-encoded data rather than human-readable plaintext. Any documents that were indexed after the license was enabled but before the restart should be re-indexed so that their stored content is encrypted.
83497	Known Issue (8.1): The AI Data Extraction template/review UI does not support non-scalar property datatypes. Do not select the following property types as extraction fields — they are string-coerced and cannot be reviewed correctly: ItemsTable (line-item tables), Reference (object reference), XML, LDAP. Password properties must never be used as extraction targets. Supported types: String, Text, Title, Email, URL, Integer, Long, Float, BigDecimal, Boolean, Date, Menu (a well-authored extraction prompt guides the LLM to a valid menu value). Full ItemsTable (line-item) extraction support is planned for a future release.
85751	On systems with Content Encryption enabled, documents added after Content Encryption was turned on may not be returned by content (full-text) searches. Property and metadata searches are not affected, and documents indexed before Content Encryption was enabled remain searchable, so the condition may not be immediately apparent. This affects systems upgraded from an earlier DocuShare release. New installations are not affected.	Add the content encryption library to the Index Server classpath: 1. Stop DocuShare. 2. In <DSHome>\config\Monitor.xml, locate the <Server> block whose <Name> is IndexServer, and append ;+DSH+\lib\cryptoContent.jar to the <StartJVMClassPath> element within that block. 3. Save the file and restart DocuShare. 4. Confirm the change took effect. The effective classpath is written to Monitor.log each time the JVM launches, and the dsix command line must now list cryptoContent.jar. 5. Only after step 4 is confirmed, re-index the affected content. The classpath change is not retroactive. Do not re-index before step 4 is confirmed. Re-indexing while content is still unreadable re-adds the affected documents with metadata only, which removes the extracted text of documents that were previously searchable.