Changelog

Reasoner now connects directly to your essential work tools, bringing live data into your indexes with full traceability. This first release focuses on project status tracking with automatic hourly updates.

• Live data connections automatically sync from your tools hourly, ensuring insights are always current
• Intelligent relationship inference identifies connections between items across different data sources
• Selective data integration lets you choose exactly which content to include in your analysis
• Unified views combine data from multiple sources into cohesive, meaningful visualizations

Currently Supported:
- Linear (tasks, projects, milestones)
- Guru (knowledge base articles)
- Google Sheets (custom data)      

Whether tracking project progress, knowledge management, or custom datasets, Reasoner Integrations give you a continuously updating view with the full traceability you expect from Reasoner.

Introducing Recall, a powerful new capability that delivers precise, contextually relevant information from your Reasoner Index with fully validated citations, transforming how you extract insights from your data sources.

• Contextual information retrieval intelligently understands your query intent and returns the most relevant information
• Comprehensive citation discovery finds relevant information across all entities in your index
• Intelligent citation validation ensures every piece of information is precisely matched to its source
• Cross-page citation matching finds relevant information even when it spans multiple pages

1
from reasoner import Reasoner
2
3
# Initialize the client
4
r = Reasoner(api_key="your_api_key")
5
6
# Run Recall against your index
7
response = r.recall("What is the capital structure?", index_uid="your_index_id")
8
9
# Access the validated citations
10
for citation in response.citations:
11
    print(f"Source: {citation.source}")
12
    print(f"Page: {citation.page}")
13
    print(f"Text: {citation.text}")

Whether you’re researching complex topics or building AI applications that require absolute accuracy, Recall delivers the precise information you need with complete confidence in its sources.

Reasoner now gives you unprecedented control over your experience with our new comprehensive settings panels, making it easy to manage everything from security preferences to team organization in one place.

• Security settings let you manage authentication preferences and access controls
• General settings for customizing your personal Reasoner experience
• Organization settings to configure workspace-level preferences
• Team management for controlling access and permissions across your organization

These new settings make it easier to tailor Reasoner to your specific needs, whether you’re working solo or managing a large team.

Building on December’s enhanced trace capabilities, Reasoner now delivers an even higher level of accuracy with our new iterative citation system that meticulously validates every reference against your source content.

• Meaningful citations include actual document names and page numbers instead of internal identifiers
• Iterative citation generation builds references through multiple validation passes for higher accuracy
• Improved reference hydration ensures citations point to the exact location in your source documents
• Enhanced error messaging clearly indicates when a trace encounters problems

These enhancements deliver more accurate, reliable citations while maintaining Reasoner’s industry-leading traceability, giving you confidence in the insights derived from your data.

Building on January’s Source Settings update, we’ve completely redesigned how you interact with individual sources in your Reasoner Index, giving you more control and deeper insights into your data.

• Source content viewer lets you examine the actual content of your documents with proper page navigation
• Detailed analysis logs show processing history with accurate entity analysis counts and processing types
• Analysis level controls allow you to initiate new analyses with different depth options when needed
• Secure source management with appropriate confirmation workflows for critical actions like deletion

These improvements make it significantly easier to understand what’s in your sources, track how they’ve been processed, and manage them effectively within your Reasoner Index.

Watch your documents transform into knowledge with Reasoner’s new source processing visualization, showing you exactly what’s happening as your documents are analyzed with real-time updates.

• Visual processing pipeline displays each stage from document preparation through entity discovery and relationship analysis
• Live progress indicators with animated elements show exactly which stage is currently active
• Detailed status messages provide specific information about what’s happening at each moment
• Entity discovery metrics show how many entities and entity types have been found as processing continues

This new capability transforms what was previously a black box into a transparent, engaging experience that helps you understand exactly how Reasoner builds its powerful knowledge indexes from your documents.

Reasoner’s entity processing pipeline has been significantly enhanced to deliver more accurate, reliable entity analysis with better relationship mapping and improved traceability.

• Reference validation ensures all citations point to actual content in your source documents
• Improved entity typing with more precise classification of discovered entities
• Smart entity filtering automatically removes low-relevance entities that appear on only a single page with no relationships
• Enhanced relationship data ensures complete information is preserved throughout processing

These enhancements deliver more accurate entity and relationship analysis while maintaining Reasoner’s industry-leading traceability, giving you confidence in the insights derived from your data.

Building on January’s Index Manager redesign, we’ve added detailed analytics, flexible configuration options, and powerful management tools that make it easier to understand and optimize your knowledge base.

• Detailed index statistics show entity counts, data keys, and relationship analyses across different processing levels
• Flexible auto-deletion policies let you set custom expiration times for indexes that are no longer needed
• Source-level analytics display entity counts, reference counts, and data keys for each individual source
• Secure deletion workflows for both indexes and sources with proper rebuilding to maintain data integrity

Whether you’re managing a single index or dozens, these new capabilities give you complete visibility and control over your data, making it easier to maintain an organized, efficient knowledge base.

Reasoner now gives you unprecedented control over how traces are generated with our new Trace Settings interface, making it easy to customize processing and integrate these settings directly into your applications.

• Visual or Query Trace selection lets you choose between different trace formats based on your needs
• Smart Follow-ups and Query Rewrite add-ons enhance your traces with additional capabilities
• Granular configuration settings with performance impact indicators help you make informed choices
• Automatic code generation creates SDK or cURL code based on your exact settings with one-click copying

This new capability bridges the gap between Reasoner’s Console and your development environment, making it simpler to build applications with the exact trace behavior you need.

Available Languages:
- Python (SDK)
- cURL (API)
Generated code includes:
- All selected trace settings
- Authentication details
- Proper formatting and imports

Reasoner now processes significantly larger documents while maintaining accuracy and traceability, expanding what’s possible with your data sources.

• Optimized entity extraction identifies key entities even in documents exceeding 300 pages
• Improved alias detection expands entity identification across your entire content library
• Refined entity typing provides more accurate classification of discovered entities
• Memory optimization allows processing of documents up to 500 pages with light analysis

These improvements mean you can now work with more complex, comprehensive documents while getting consistent, reliable results from Reasoner’s traceability engine.