Skip to content

Release Notes FT Intelligence

Release 3.0

Released on: April 8th, 2026

  • Admin: Improved UI and responsiveness across all views.
  • Admin: Added new "Settings" tab for centralized configuration.
  • Admin: Added new "Examples" tab with support for user, parent account, and default prompt sections.
  • Admin: Added document exclusion and inclusion controls for managing document visibility.
  • Admin: Added support for CSV file uploads and increased file size limit to 100MB.
  • Admin: Enhanced uploads management with separate subproject and parent project sections.
  • Admin: Improved document status tracking, filtering, and validation.
  • Admin: Improved data loading, token refresh, and error handling.

  • Admin: Fixed document duplication, upload permission, and write-access issues.

  • User: Major UI/UX overhaul with improved layout, styling, and responsiveness across all components.

  • User: Added virtual scrolling for chat log entries for improved performance.
  • User: Added chat renaming functionality and context menu for chat log actions.
  • User: Added confirmation popup on chat deletion.
  • User: Added agent graph mode selector for choosing between agent types.
  • User: Added document visibility management with hide upload functionality.
  • User: Enhanced chat input with tool suggestions and document attachments.
  • User: Added tooltips and metadata labels to tools for improved discoverability and guidance.
  • User: Improved real-time streaming with proper state handling for thinking indicators and response traces.

  • User: Fixed issues with socket connections, queued questions, and session switching.

  • Backend: Added event handling for chatbot to send post messages to the platform.

  • Backend: Added retry/backoff strategy on FieldTwin API calls for improved reliability.
  • Backend: Added per-user rate limiting on the /ask endpoint (configurable, credit-based).
  • Backend: Added chat history retention policy (session-based).
  • Backend: Added support for searching documents uploaded in linked projects.
  • Backend: Improved PromptGuard scope enforcement with dialog-aware message handling.
  • Backend: Improved mid-stream error handling with frontend notification.
  • Backend: Optimized token counting using async worker thread with wasm tiktoken.
  • Backend: Improved database connection management using connection pooling.
  • Backend: Improved Kubernetes job error handling and non-blocking HTTP responses.
  • Backend: Improved tool category selection for project assets and connections.
  • Backend: Fixed parent/child integration settings for document propagation.

  • Backend: Fixed upload failures for certain doc/docx/pdf files.

  • General: Added Langfuse observability deployment configuration.

  • General: Added integration documentation and tool capabilities reference.
  • General: Improved deployment strategy and Helm chart configuration.
  • General: Updated LangChain, Langfuse, and other core dependencies.
  • General: Fixed Node module vulnerabilities.

Release 2.0

Released on: September 17th, 2025

This initial release introduces FieldTwin Intelligence, your AI assistant for FieldTwin that enables interactive conversations with your FieldTwin data. Powered by Large Language Models (LLM) and implementing Retrieval-Augmented Generation (RAG) with Autogen Agents using LangGraph, this comprehensive solution brings immediate value through rapid knowledge retrieval, design validation, and intelligent workflow automation.

  • Admin: Added comprehensive management panel for creating and configuring example queries available to users.
  • Admin: Added centralized document repository to manage files available across all projects within the account.
  • Admin: Added multi-format support for uploading and managing PDF, DOCX, DOC, TXT, and MD files.
  • Admin: Added real-time status tracking for document processing and indexing.
  • Admin: Added upload history tracking with timestamps and file lifecycle management.
  • User: Added natural language processing for typing questions and commands in plain English.
  • User: Added interactive response actions with regenerate, feedback, and response details options.
  • User: Added real-time visual feedback when AI executes tools or accesses data sources.
  • User: Added conversation flow visualization showing AI reasoning process and tool usage.
  • User: Added searchable chat history with temporal organization (7 days, 30 days, older sessions).
  • User: Added context preservation within sessions for intelligent follow-up questions.
  • User: Added multi-source access to User Guides, API documentation, and uploaded files.
  • User: Added visual source indicators with clear badges showing active and available data sources.
  • User: Added multi-format file upload support for PDF, MD, DOCX, TXT, DOC, CSV, and JSON files.
  • User: Added processing status tracking with clear indicators (Ready, Process).
  • User: Added ready-to-use query examples that can be executed immediately.
  • User: Added responsive design that adapts seamlessly to different screen sizes and devices.
  • User: Added asset discovery to search and browse account-level asset definitions, connections, wells, shapes, and layers.
  • User: Added metadata management to query and search metadata definitions by name, definition ID, or vendor ID.
  • User: Added advanced asset querying with support for category and metadata filters.
  • User: Added connection analysis for querying project connections with filtering capabilities.
  • User: Added well management to search and analyze project wells with comprehensive metadata-based filtering.
  • User: Added bulk update operations for updating properties of multiple assets, connections, wells, shapes, or layers.
  • User: Added asset creation capability to create new staged assets in projects using account asset definitions.
  • User: Added CSV generation to export project data to downloadable CSV files with custom formatting.
  • User: Added chart creation to generate visual charts from project data using Chart.js integration.
  • User: Added DXF export to create DXF files from connection geometry for CAD integration.
  • User: Added semantic search with AI-powered similarity search across documentation and uploaded files.
  • User: Added fuzzy matching with intelligent name and category matching with typo tolerance.
  • User: Added Model Context Protocol (MCP) integration for standardized tool access in external systems.

This comprehensive release establishes FieldTwin Intelligence as a powerful AI assistant that transforms how users interact with FieldTwin data, providing immediate value through rapid knowledge retrieval, intelligent analysis, and automated workflows.