Datafari 6.2 Released with Simplified AI Integration and Vector Search
Table of Contents
The latest version of the open-source enterprise search solution boasts improved generative AI capabilities, vector search, and a new RAG module.
The latest iteration of Datafari, version 6.2, has been launched, featuring streamlined integration of third-party generative AI tools. This enhancement was partially funded by the European project “NGI Search Neural Datafari,” wich also spurred vector search advancements within the Apache solr project.
Datafari serves as a business research solution designed to enable employees to efficiently locate relevant data, nonetheless of its format or location. The system indexes data and documents from diverse sources and file formats, allowing users to search for documents and leverage generative AI for advanced querying.
This release emphasizes the free and open-source version, although the proprietary business version also includes new functionalities.
Key updates and changes since Datafari 5.3 include:
- Addition of a Retrieval Augmented generation (RAG) module.
- Migration to solr 9.8 with vector search capabilities.
- Implementation of an LLM (Large Language Model) call module for indexing.
- Automated management of indexed document chunking.
- Creation of an autonomous analytics module to replace Zeppelin for optimized resource utilization.
- Introduction of a Regex Connector for indexing.
- Technical overhaul of the graphical interface using React.
- Transition to V2 of DataFari REST APIs.
- Autonomous agent prototype for localizing AI models.
- General bug fixes.
A demonstration video is available to showcase the new features. A Swift start Guide and thorough documentation are also provided, covering usage, management, and growth aspects. The AI modules require a compatible AI model server. Datafari does not host AI locally, instead connecting to services like Openai or locally hosted models.
What Can datafari Do?
Datafari functions as an AI-enhanced enterprise search engine, designed to unify knowledge management by indexing and analyzing organizational documents and metadata from multiple sources and formats, while managing security and administration.
Admin-Side Features (Free Version):
- Administer connectors to various data sources using Apache Manifoldcf connectors, including SharePoint, Confluence, Alfresco, and file shares.
- manage the relevance algorithm for ranking documents in search results.
- Activate vector and RAG search modules.
- highlight specific documents for designated queries.
- Create and manage user roles.
- Access tool usage statistics.
- Create promoted links (similar to Google Adwords).
- Manage synonyms.
- Access additional features via the Confluence documentation.
User-Side Features:
- Perform simple and advanced searches.
- Preview search results.
- Interact with documents using the RAG module.
- Utilize spelling correction and auto-completion.
- Filter results using facets.
- Save results to a favorites basket.
- Create email alerts for new or modified documents matching a query.
“Datafari is a search engine for company enriched at AI: family members of knowledge management tools, research solutions federate knowledge…”
Future Enhancements
The development team plans to introduce more generative AI-based features throughout 2025.
Feedback and suggestions are encouraged to further improve the product. users are also encouraged to share their experiences online.
Frequently Asked Questions
- What is Datafari?
- Datafari is an open-source enterprise search engine that uses AI to help organizations find and manage their internal knowledge.
- What are the key features of Datafari 6.2?
- Key features include simplified AI integration,vector search,a RAG module,and improved administration tools.
- How can Datafari improve my organization’s productivity?
- By providing a unified search interface and leveraging AI, Datafari helps employees quickly find the information they need, improving decision-making and reducing wasted time.
