Changelog

August 2026 Platform Update

Version 3.0.0 · August 24, 2026

Introducing the Solstice Help Center

We've introduced a new Help Center directly within Solstice, giving users a centralized place to stay informed, access documentation, and connect with the CelestiQ support team.

• Updates – View product releases, maintenance notes, and recent changes across the Solstice platform.
• Documentation – Access training materials, usage guides, and tenant-specific instructions for your organization.
• Contact Support – Send messages directly to the CelestiQ support team whenever you need assistance.
• Easy Access – Open the Help Center using the Help button in the bottom-right corner of the chat or from the Settings dropdown menu.

Solstice Desktop App is Now Available

Solstice is now available as a native desktop application for both Windows and macOS, bringing the Solstice experience directly to your desktop.

• Windows & macOS – Download the new Solstice desktop application for your supported device.
• Download Center – Desktop applications are available at solstice.celestiq.ai/downloads.
• In-App Downloads – When enabled for your organization, the Downloads page can also be accessed directly from the Settings dropdown menu.
• Mobile Apps Coming Soon – Native mobile applications are in development and are expected to be released in the coming weeks.

Next-Generation Voice Experience

We've upgraded our text-to-speech models with a significantly more natural and expressive voice engine, making spoken responses feel more fluid across languages and use cases.

• More Natural Speech – Upgraded voice models deliver smoother pacing, improved pronunciation, and more natural-sounding responses.
• 70+ Languages – The new multilingual voice engine supports more than 70 languages for a significantly broader global experience.
• Improved Multilingual Quality – Better handling of language-specific pronunciation, cadence, and conversational speech.

Smarter File Generation & Live Simulations

We've continued to expand Solstice's creation capabilities with more intelligent design, richer generated outputs, and an entirely new ability to create interactive experiences directly inside conversations.

• More Intelligent Design – Generated files now use improved layout, structure, visual hierarchy, and formatting to produce more polished results.
• Live Simulations in Chat – Solstice can now create interactive simulations that run directly within the chat experience.
• Improved Papers & Documents – Continued advancements in generating structured papers, reports, and other long-form content.
• Enhanced Tables – Smarter generation and presentation of structured data and tables.
• Image Creation – Ongoing improvements to visual generation and image-based workflows.
• Presentation Generation – Continued improvements to presentation design, structure, and generated slide content.

These capabilities, particularly live simulations and advanced file creation, remain a work in progress and will continue to improve rapidly. We encourage you to provide feedback directly in chat or contact the CelestiQ team through the new Help Center.

Personalization & Interface Enhancements

This release includes a broad set of UI and UX improvements designed to make Solstice more customizable, accessible, and efficient across devices.

• Redesigned Sidebar – Reworked navigation and sidebar behavior provide a cleaner and more streamlined experience.
• Additional Fonts – New typography options include CelestiQ Sans, CelestiQ Serif, OpenDyslexic, and SF Pro on supported Apple devices.
• Chat Text Customization – Adjust chat text size and weight to better match your reading preferences.
• Additional Theme Colors – Expanded appearance options provide more ways to personalize the Solstice interface.
• Delete Account In-App – Eligible users can now delete their account directly from within the application.
• Settings Improvements – Additional navigation and organization refinements make important account and application options easier to find.

Additional Improvements & Refinements

This release also includes numerous bug fixes, interface refinements, performance optimizations, reliability improvements, and additional enhancements throughout Solstice. We're continuing to iterate quickly across chat, file creation, voice, desktop, mobile, and the broader platform experience based on user feedback.

July 2026 Platform Update

Version 3.0.0 · July 27, 2026

Welcome to Solstice

CelestiQ's platform is evolving into Solstice, the next generation of our AI platform, reflecting a unified intelligence platform built for faster research, smarter automation, and a more seamless user experience on all devices.

• Introducing Solstice – Our new platform identity that represents the future of enterprise AI.
• Desktop & Mobile Applications – Native Solstice desktop and mobile apps are actively under development and are expected to launch next month.
• Early Beta Access – Contact your administrator to enroll in the Solstice Beta program beginning in early August.

Application Expert is the New Standard

Application Expert has become the default experience for every new conversation, transforming into a true all-inclusive AI assistant capable of intelligently selecting the best tools and knowledge sources for every request.

• Unified Intelligence – Application Expert now has access to every data source previously available across all specialized modes.
• Automatic Tool Selection – The agent determines which search pipelines, retrieval systems, and services are required without manual mode selection.
• Focused Modes Still Available – Specialized chat modes remain available whenever targeted searches are preferred.

Redesigned Application Expert Topics

We've completely reworked how Application Expert approaches complex engineering and business requests.

• Cross Referencing – Automatically compare information across multiple documents, systems, and sources.
• Part Lookup – Faster identification and validation of electronic components and technical references.
• Company Research – Expanded organization intelligence and business research capabilities.
• Pricing Research – Improved pricing discovery, comparison, and sourcing workflows.
• Dynamic Workflow Planning – Instead of fixed search paths, the agent determines which tools are required to solve each task.

Next-Generation Query Engine

One of the largest backend upgrades to date introduces a completely redesigned search architecture capable of handling significantly more complex requests.

• Multi-Pipeline Search – Multiple retrieval pipelines can execute simultaneously from a single request.
• Intent-Aware Processing – Queries are analyzed before execution to determine the optimal search strategy.
• Parallel Knowledge Retrieval – Internal knowledge, external web search, and specialized retrieval systems work together automatically.
• Smarter Query Routing – Requests are dynamically routed to the services best suited for the task.
• Improved Search Relevance – Better ranking and result aggregation produce faster and more accurate answers.

File Generation Improvements

Document creation has become significantly more reliable, with continued improvements planned for future releases.

• Fixed File Creation Displays – Generated documents now display correctly after creation.
• Improved Generation Reliability – More dependable file generation workflows across supported formats.
• Continuing Improvements – Ongoing enhancements to file quality, formatting, and export capabilities remain in active development.

Performance & Reliability Improvements

This release includes a number of infrastructure upgrades designed to make Solstice faster and more dependable.

• Knowledge Base Stability – Fixed timeout issues when querying large knowledge collections.
• Improved Web Search Reliability – Resolved timeout issues affecting external search requests.
• Conversation Startup Fixes – Fixed an issue where some new conversations could begin with an "aborted" message.
• Backend Optimizations – Additional infrastructure improvements for faster, more reliable responses.

Additional Improvements & Refinements

This release also includes numerous bug fixes, search optimizations, infrastructure enhancements, UI refinements, and ongoing performance improvements across the Solstice platform as we prepare for the upcoming desktop and mobile releases.

May 2026 Platform Update

Version 2.3.2 · May 29, 2026

Enhanced File Uploads & Document Intelligence

Working with files inside conversations is now significantly more powerful. Agents can analyze larger collections of documents, maintain context across uploads, and deliver more accurate responses from complex content.

• Universal File Support – Upload virtually any readable file type including PDFs, spreadsheets, images, presentations, technical documents, and more.
• Up to 10 Attachments Per Message – Provide multiple files simultaneously for richer context and cross-document analysis.
• Persistent File Context – Uploaded files remain available throughout the conversation, allowing Agents to continuously reference previously uploaded materials.
• Improved File Understanding – Enhanced extraction, document parsing, table recognition, and contextual reasoning improve accuracy when analyzing complex content.
• Better Structured Data Extraction – Improved handling of spreadsheets, reports, tables, and technical documentation.
• Attachment Experience Improvements – Refined attachment previews, upload workflows, and file management throughout the chat experience.

Knowledge Ingestion & Retrieval Improvements

Significant improvements have been made to enterprise knowledge ingestion, helping organizations build higher-quality knowledge bases and retrieve more accurate information from their content.

• Enhanced URL Discovery & Tracking – Improved website crawling, monitoring, and synchronization for customer knowledge sources.
• More Accurate Fact Extraction – Better identification and preservation of important business information during ingestion.
• Catalog & Datasheet Optimization – Specialized processing for product catalogs, technical datasheets, and structured product information.
• Improved Data Cleaning Pipelines – Cleaner source data results in more reliable retrieval and higher-quality responses.
• Enhanced Retrieval Relevance – Faster and more precise matching between user questions and organizational knowledge.

Evaluations & Quality Intelligence

Introducing Evaluations, a new layer of quality monitoring designed to help organizations better understand Agent performance, response quality, and overall user experience.

• Automated AI Grading – Every conversation can now be evaluated automatically using our built-in quality assessment system.
• Transcript-Level Evaluations – Review evaluation results directly alongside conversations within transcripts.
• Dashboard Evaluation Analytics – Track quality trends, response effectiveness, and user experience metrics across your organization.
• Performance Visibility – Identify opportunities for optimization, training, and knowledge improvements with greater confidence.

Executive Dashboard Overhaul

The Executive Dashboard has received a substantial redesign focused on usability, scalability, and deeper operational visibility.

• Full-Screen Dashboard Experience – Dedicated pages provide more room for analytics, administration, and reporting.
• Dashboard Refresh – Cleaner layouts, improved navigation, and a more modern visual experience throughout the platform.
• Enhanced Analytics UI/UX – Improved workflows for exploring usage, adoption, engagement, and evaluation metrics.
• Integrated Evaluation Insights – Evaluation data is now available directly within reporting and analytics workflows.
• Mobile Optimization – Significantly improved responsiveness and usability across phones and tablets.

GPT-Image-2 Integration

Image generation capabilities have been upgraded to GPT-Image-2, delivering higher-quality visual outputs and improved instruction following.

• Higher Fidelity Outputs – Improved image quality, consistency, and detail across generations.
• Better Prompt Adherence – More accurate interpretation of user instructions and visual requirements.
• Enhanced Diagram & Illustration Generation – Better results for technical, business, and creative workflows.

Performance & User Experience Improvements

This release includes dozens of reliability, usability, and performance enhancements throughout the platform.

Chat Experience

• More Reliable Streaming – Improved response generation stability during long-running conversations.
• Background Message Processing – Responses continue generating even when navigating away from the active chat.
• Smoother Response Rendering – Reduced visual jitter and improved generation animations for a more natural experience.
• Cleaner Chat Performance – Faster interactions and improved responsiveness across extended conversations.

Interface Improvements

• Copyable Tables – Structured table outputs can now be copied directly into spreadsheets, documents, and reports.
• Improved Mathematical Formatting – Cleaner rendering and presentation of equations and technical content.
• Enhanced Code Blocks – Better readability, formatting, and developer-focused presentation.
• Refined Attachment Modal – Improved file previews, upload feedback, and attachment management workflows.

Additional Improvements & Refinements

This release also includes numerous bug fixes, stability improvements, infrastructure upgrades, UI polish, and performance optimizations across both the Agent platform and Executive Dashboard.

March 2026 Platform Update

Version 2.3.1 · March 25, 2026

Persistent Chat Sessions

Your conversations are now more flexible and continuous than ever.

• Independent Chat Sessions – Each chat now operates as its own persistent session, allowing you to revisit and continue discussions at any time.
• Seamless Topic Continuity – Pick up right where you left off without losing context.
• Legacy Chat Support – Previous chats remain visible, but cannot be continued as active sessions.

Personalized User Preferences

We’re introducing deeper customization to tailor your Agent experience.

• Custom Instructions – Define how your Agent communicates and behaves to better match your preferences.
• Profile-Based Context – Your name, language, and company are now automatically included in every new chat session.
• Persistent Personalization – Instructions are applied across all future conversations for a consistent experience.

Executive Dashboard Enhancements

Improved visibility and analytics for leadership and management teams.

• Mode Visibility in Transcripts – Every message now includes its associated mode for clearer insight into system behavior.
• Analytics UI Overhaul – A redesigned analytics experience with more interactive tools for monitoring usage, performance, and engagement trends.

Live Status & System Notifications

Stay informed in real time about platform health and updates.

• Live Status Banners – Receive instant notifications across both the platform and dashboard during maintenance or service events.
• Status Subscription Available – Subscribe for updates at status.celestiq.ai.

Feature Sunset Notice

As part of ongoing platform consolidation and intelligence upgrades, the following features are being deprecated.

• Web Search Button Removed – Native web search functionality has been sunset as its capabilities are now fully integrated into Agent responsibilities.
• Product Search (Distribution & Manufacturer Rep Models) Sunset – These tools have been absorbed into the enhanced Application Expert system for a more unified experience.

Additional Improvements & Refinements

This release includes a wide range of usability, performance, and intelligence upgrades.

Platform UI/UX Enhancements

• Agent Thinking Animation – Visual feedback while the Agent processes responses.
• Improved Streaming Experience – Smoother, more natural response rendering.
• Enhanced Code & Math Rendering – Cleaner, more readable formatting for technical outputs.
• Scroll-to-Bottom Shortcut – Quickly return to the latest message in long conversations.
• Session Mode Memory – Chats now remember the last mode used and automatically restore it.

Application Expert Knowledge Base

• Improved Retrieval & Ingestion Pipelines – Faster, more accurate knowledge access.
• Expanded File Support – Larger file uploads now supported.
• Refined Data Cleaning Systems – Higher-quality structured outputs.
• Enhanced AI OCR Pipeline – Better extraction from complex documents.
• Product Catalog Integration – Structured product data can now be ingested directly into the system.

Language & Localization

• Multi-Language Support Expanded – Select your preferred language for both the platform and Agent.
• Accessible Language Controls – Available on the login screen and within Profile Settings.

January 2026 Platform Update

Version 2.3.0 · January 31, 2026

Upgraded Application Expert Mode

All Agent’s Application Expert Mode has been fully transformed into one of the most powerful proprietary engineering intelligence systems in the industry.

• Automated Data Organizing & Formulation – Instantly structures complex technical information into actionable intelligence.
• Self-Learning Feedback Engine – A new automated learning system continuously improves the agent through real-world user feedback.
• Always Growing Intelligence – The platform is now constantly refining its knowledge depth and response accuracy.

Executive Transparency Dashboard

We’re introducing a brand new CelestiQ Dashboard, built for executive and technical leadership teams.

• Full Database Visibility – View all ingested documents, web files, and structured engineering data in one secure system.
• Document Upload & Management – Users can upload and handle technical files with complete transparency.
• Transcript & Feedback Analytics – Track conversations, user feedback, and advanced usage insights across the organization.

Secure Citations & Private Knowledge Storage

All knowledge sources are now fully traceable.

• Every File is Cited – Responses include secure citations from stored documents and web sources.
• Private Database Protection – Sensitive engineering data is stored safely in CelestiQ’s secure infrastructure.

Voice & Speech Recognition Now Live

Your Agent now supports voice-driven workflows.

• Speech Recognition Added – Activate voice input directly beside the Send button.
• Read-Aloud Messages – Every response can now be played back aloud.
• Custom Voices Available – Change voices anytime in the Settings → General tab.

For advanced voice capabilities, contact the CelestiQ service team.

Additional Improvements & Refinements

This release also includes major UI polish and platform stability upgrades across the system.

• Personalized Typography – Your Agent now automatically adapts fonts based on your operating system, using Inter or SF Pro.
• Sidebar UI Refresh – Navigation and workspace organization have been visually upgraded for a cleaner experience.
• Enhanced Transcript PDF Generation – Transcript exports are now more polished, branded, and reliable.
• Version Sync Bug Fix – Resolved an issue where some users could become stuck on outdated platform versions.

October 2025 Platform Update

Version 2.2.0 · October 10, 2025

Introducing QuantiQ Search

QuantiQ Search is here — CelestiQ’s new premier mode that brings the global electronics market directly into your workspace.

• Universal Access – Instantly scan live pricing, stock, and technical data from distributor networks worldwide.
• Real-Time Intelligence – Compare parts, verify specs, and analyze availability with data that updates as you search.
• Seamless Integration – Works alongside all company-specific modes to deliver a unified sourcing experience.

Folder Organization for Chats

Stay organized with new Chat Folders — group related conversations together for better project management.

• Move Chats to Folders – Easily assign chats into folders for streamlined organization.
• Download Folder Transcripts – Export all conversations within a folder as a single, branded transcript for reporting or archiving.

Improved Mobile Experience

The CelestiQ interface has been redesigned for enhanced mobile usability, ensuring full access and performance on any device.

• Optimized Layout – Responsive design adapts seamlessly for smaller screens.
• Full Functionality on the Go – Access, organize, and download chats directly from your phone or tablet.

August 2025 Platform Update

Version 2.1.2 · August 12, 2025

What’s New in This Release

Live Streaming Output

Agents now deliver answers token-by-token in real time.

• Faster Insights – See results as they’re generated, cutting wait time for long or complex queries.
• Improved Engagement – Continuous feedback keeps you confident the search is working.
• Efficiency Boost – Streaming reduces the time to receive a complete answer by up to 75%.

Processing Animations

Added dynamic animations for Web Search, File Uploads, Image Analysis, and Reasoning Mode — making it clear when your request is in progress.

Other Improvements

• Smoother real-time updates across all modes
• Better performance for large queries
• UI refinements for a cleaner experience

June 2025 Platform Update – Part 2

Version 2.1.1 · June 19, 2025

What’s New in This Release

Chat Transcript Downloads

You can now download transcripts as PDFs or export your entire chat history by session. Access this from the Sidebar menu for individual downloads or use the Data tab under Settings to export everything in one click.

New Settings Experience

The Settings panel has been completely revamped with four intuitive tabs:

• General – Manage appearance and core behavior.
• Profile – View and update your name, email, company, or reset your password.
• About – See agent name, version, changelog, platform instructions, terms & policies, and live service status.
• Data – Export all chats, delete conversations, or reset app data.

The Service Status page now shows live info for any disruptions across CelestiQ’s providers (OpenAI, Anthropic, AWS, database/server).

Knowledgebase Awareness

When operating in General Use mode, a warning badge appears in the top-left next to the model selector, letting you know the custom knowledgebase is not being used.

Left Click = New Tab

By extremely popular demand (and a whisper of internet rage), left-clicking links in chat now opens them in a new tab. That’s right — no more Cmd+click acrobatics or mysterious vanishing sessions!

Additional Improvements

• Streamlined chat export flow and data handling
• Improved Settings layout for faster navigation
• More transparency around platform health and uptime
• Clearer cues for knowledgebase vs general assistant use

June 2025 Platform Update

Version 2.1.0 · June 6, 2025

What’s New in This Release

PDF Uploads

Upload up to 3 PDFs per message. Each appears as a clickable, previewable bubble that remains available during your session. Files are not accessible after a page refresh but their previews remain for context — ideal for spec sheets, BOMs, and application documents.

Image Uploads

Supported formats include .jpg, .jpeg, .png, and .gif. Images are rendered inline, zoomable, and persist even after page refresh. You can now also paste images using Ctrl+V for faster visual feedback loops.

Use the + button or Ctrl+V to upload PDFs or images.

Web Browsing Mode

Toggle by clicking the globe icon before sending a message. Enter a URL or a search query to pull live data from the web. Mode resets after each message.

Reasoning Mode

Click the brain icon to activate OpenAI’s o3 model for deep reasoning and multi-part analysis. Perfect for technical problem solving and long-context understanding. Automatically turns off after one use.

General Use Mode

A standalone experience for everyday and advanced AI tasks. Seamlessly move between freeform queries and technical engineering help without switching sessions.

Combine Modes for Smarter Use

• Web + Application Expert — Combine live internet data with private company knowledge.
• Reasoning + Application Expert — Analytical problem solving grounded in your internal models.
• Reasoning + Web + Application Expert — Deep research, competitive insights, and informed design tradeoffs.
• File Upload + Application Expert — Extract spec data or BOM content to drive intelligent responses.
• File + Reasoning + Application Expert — Document analysis and spec interpretation with domain intelligence.
• General + Web — Explore topics using real-time data in a flexible assistant experience.
• General + Reasoning — Layer general questions with critical thinking and synthesis.
• General + Web + Reasoning — All-in-one for exploration, planning, and multi-dimensional research.
• General + File + Reasoning — Analyze technical files and reports within a broader task context.
• General + File — Use uploaded documents in casual or structured queries without losing flow.

Other Improvements

• Updated UI for smoother uploads and cleaner button styles
• Preview PDFs and images inline
• Support for latest OpenAI (o3) and Anthropic Claude models
• Expanded message input and output limits
• New image paste support via Ctrl+V
• Performance enhancements and various bug fixes