New Features

Image Generation

• Selecting an image generating model will automatically generate an image.
• Added openAI's DALL-E-2 & DALL-E-3 and gpt-image-1, which is OpenAI's latest and best image generating model, corresponding to gpt-4o in the chatGPT app. Gpt-image-1 is capable of editing images as well.
• Added Imagen-3.0 from Google. Gemini-2.0-flash-experimental is currently not available in Europe, so we are not able to provide it for the time time being
• Added Grok-2-image from XAi

Incognito Mode

• Added an incognito profile. Chats in the incognito profile will be deleted if you change to another profile. You can keep your incognito chat by changing the chat's profile.

Hotkeys for Chat

• Added keyboard shortcuts for common chat actions

Message Encryption

• Slowly introducing encryption and decryption of prompts and responses in messages, more to come later with this.

Model Updates

Grok-3 Integration

• added Grok-3, which is XAI's latest text model
• OpenAI's o1-mini model will soon be deprecated, so it is removed from our options.

UI Improvements

Chat Interface Enhancements

• Now hiding resize grip when chat is at maximum width
• Improved overall chat experience

Other Changes

Subscription Changes

• Updated subscription plans and pricing structure

Llama4 models, chat profiles, enhanced text copying

New Models

We have added Meta's latest Llama 4 based models. The following models have been depreciated and have been removed from our service:

• llama-3.2-1b-preview
• llama-3.2-3b-preview

Chat profiles

We've added a "Profiles" feature to help you organize your chats. The current profile is visible in the top left corner of the application.

Enhanced text copying

We have added a click based way to copy blocks of text, 1 click copies the current word, 2 cliks the current sentence, 3 clicks the current paragraph and 4 quick consecutive clicks the whole response. This way of copying is disabled in code blocks.

Patch Updates

Image Interaction Modal

• Images now open in their own modal when clicked on them

What's Next?

We hope these updates enhance your experience and provide you with more powerful tools to manage your interactions. We are currently working on getting image generation out hopefully by the end of this week

Web Search, Grok Integration, Gemini 2.5, Image uploading

New Model Integrations

We've expanded our AI model ecosystem with two significant additions:

Grok Integration: We're thrilled to welcome xAI's Grok to our family of supported models. We will be adding more of Grok's models during the coming weeks.

Gemini 2.5 Support: Google's latest and most performative Gemini 2.5 model is now available on Octofy, bringing its enhanced capabilities to our platform. This cutting-edge model offers improved reasoning and understanding, giving you even more options for your AI interactions.

Enhanced Capabilities

Web Search Integration: All models on Octofy can now access the web through our Google Custom Search API integration. This feature allows our AI models to retrieve and display current information from the internet, making responses more informative and up-to-date than ever before. Links to used sources are also posted on responses that use the web integration.

Extended Thinking Mode for Claude 3.7 Sonnet: We've implemented an extended thinking mode specifically for the Claude 3.7 Sonnet model. This may result in slightly longer response times but with more in-depth answers.

Image Support in Prompts: Image can now be used as part of your prompts. The following models support using images: Claude 3.5 Sonnet, Claude 3.7 Sonnet, Claude 3.5 Haiku, Claude 3 Opus, GPT o1, GPT 4o, GPT 4o mini, GPT 4.5, Gemini 2.0 flash, Gemini 1.5 Pro, Gemini 1.5 Flash. You can also use images in our automatic selection.

What's Next?

This update represents our ongoing commitment to making Octofy the most versatile and powerful AI platform available. We're constantly working on new features and improvements based on your feedback.

Updated Message Limits and Sign-Up Flow

Updated message quotas

• Free tier: 100 light model messages per month
• Pro subscription: 500 premium messages and 25 ultra model messages per month

These limits are based on observed usage patterns. Based on user behavior, most users didn’t reach the previous message cap. This adjustment allows us to increase the context window and improve model performance per message. We’ll monitor usage and expand these limits if needed, especially for users with heavier workloads.

Sign-Up Flow Changes

The sign-up and sign-in process now uses Clerk’s native flow again.

Preferred Model Defaults and Thought Process Settings

Preferred Model Logic

Octofy is now automatically set as the default model for users without a preferred model. If a user deselects Octofy, the next available model is suggested instead. We’ve also resolved an issue where updating preferences could cause an infinite loop in the settings panel.

Thought Flow Visibility

Users can now choose to hide DeepSeek model's thought process via a new setting. This option is available in preferences.

Update in subscription models

Simplifying our subscription offerings, we have removed our light and ultra subscription tiers. In the future there will be only one subscription tier with unlimited free messages, 1 000 premium and 50 ultra messages for the same price as our previous "Pro" subscription.

New models

Our model lineup now includes OpenAI's latest additions: GPT-O1, GPT-O1-mini, GPT-O3-mini, and GPT-4.5-preview-2025-02-27.

Updates in response rendering

The are more and more models that have a "thinking process" before giving the main answer. We have implemented a preliminary way of displaying this data. We want to provide this data but we do not want it to be too much in the way. At the moment there is an option to opening and closing the thinking process, this is at the moment implemented for asnwers generated with Deepseek R1. In addition, we have added rendering of mathematical formulas in our chat responses

Hotkeys

We have added simple hotkey implementation. Currently there are hotkeys for navigating (new chat, settings, search) and for toggling the sidebar. More hotkeys are to come during the following weeks.

New OpenAI Models and Model Improvements

New Models and Providers

We have added OpenAI’s models: GPT-O1, GPT-O1-mini, GPT-O3-mini, and GPT-4.5-preview. These models bring improvements in speed, reasoning, and creative output.
DeepSeek has also been added as a new model provider, expanding the range of supported AI technologies.

Improved Model Naming

Model names are now displayed in a clear and consistent format across the entire app. This includes support for special cases like Claude, GPT, LLaMa, and DeepSeek models.

Thought Process Enhancements

The thought process in DeepSeek is now visually separated in the UI using blockquote styling for easier readability and better distinction from the model’s final answer.

Design and UI Improvements

• Font size and type preferences now apply globally across the site
• Font rendering has been improved for cross-browser consistency
• Manifest and PWA setup updated for more reliable offline support

Improved Math Rendering, Accessibility, and Copy Features

Math and Copy Features

Math formulas are now rendered in model responses. Copy buttons have been refined: they appear only when content is ready, support table/text copying, and behave consistently on mobile.

Accessibility Feedback

Users can now report accessibility issues directly via a new feedback option.

UI Fixes

• Top bar layout improved for mobile and desktop
• Prompt formatting and resizable chat fixed
• Tooltip and button interactions adjusted for mobile
• Visual consistency tweaks across chat and settings

Settings and Hotkeys

• Settings menu updated to dropdown style
• New hotkeys modal added with shortcuts for core actions

Signup Flow

Signup now starts with email to streamline the experience.

PWA Enhancements

Offline detection and feedback now work more reliably.

Our beta version is now online!

With our beta release comes a completely updated and fresh landing page.

New models

We have added Anthropic's newest model Claude 3.7 sonnet and 70b distilled version of Deepseek's R1 model. The Deepseek model is provided via Groq. We will be providing access to the original Deepseek model as soon as they open their API for new customers.

New settings

We have added a subscriptions tab to the settings for mor easier subscription management. In addition we have added the possiblity to edit the chat width.

Changes in the progressive web app (PWA)

We have improved PWA offline functionality by properly configuring the service worker. This enables the app to work offline as a standalone PWA even when the server is not running. We have also added an offline page fallback for navigation requests

Enhanced Search and Model Information

Search Functionality Overhaul

• Improved filtering by date and starred/unstarred conversations
• Better search performance and responsiveness
• Added loading states for smoother experience

Model Information

• Added release and update dates for AI models
• Better model information display in the interface
• Improved model selection experience

Response Rendering

• Improved markdown and plaintext rendering in code blocks

Backend Improvements

• Various performance optimizations
• Enhanced data synchronization

Improved Sidebar and Subscription Features

Sidebar Improvements

• Dynamic height calculation based on viewport
• Smart distribution between starred and unstarred chats
• Better support for vertical monitors
• Improved space allocation for different screen sizes

Model Selection

• New data table component for model selection
• Save selected models to your profile
• Improved model name formatting
• Enhanced filtering and selection logic

Subscription System

• Added monthly message quotas for different subscription tiers
• Created light version of recommendation function for free users
• Added alerts for monthly quota usage
• Improved subscription type management

Backend Improvements

• Various performance optimizations
• Enhanced system stability

These improvements are part of our ongoing commitment to making Octofy more powerful, user-friendly, and reliable. We're continuously working to enhance your experience!

Model Sync and Chart Improvements

Model Synchronization

• Chosen model now syncs with preferred model when updated
• Selected model correctly reflected in chat interface
• Model selection still persists across chats

Chart Visualization

• Fixed chart labels display with many categories
• Added Safari support for image downloads
• Improved mobile experience

Model Updates

• Updated available models for better performance
• Improved token handling for various models

Settings Interface Redesign

Settings Interface

• New tab container style component for consistent design
• Updated Account page to display subscription type
• Improved Profile page layout
• Better organization of settings sections

Preferences UI

• Added visual previews for light/dark mode themes
• Improved mobile responsiveness for keyboard shortcuts
• Enhanced accessibility with proper labeling
• Better section organization

System Improvements

• Various performance enhancements
• Improved data handling

Analytics

• Added usage analytics
• Helps identify areas for improvement

Visual Design Refresh

Visual Design

• New color palette based on design system
• Consistent component styling
• Documented color usage
• Refined sidebar and border colors

Theme Support

• Added light/dark mode switcher on landing page
• Improved color contrast in both modes
• Better theme adaptation for all components

UI Components

• Updated button component with purpose-specific variants
• Replaced Montserrat with Inter font for better readability

These visual improvements create a more cohesive, accessible, and pleasant experience across the entire application. We're continuously working to make Octofy not just powerful, but also beautiful and easy to use!

Starred Conversations and Settings Tabs

Starred Conversations

• Mark important chats with a star
• View starred conversations in a dedicated section
• Toggle star status from the chat list

Settings Interface

• New tabbed navigation system
• Dedicated "Prompt Preferences" tab for AI model settings
• Model suggestion dialog in Models tab
• Improved layout and visual design

Theme Updates

• Switched from Montserrat to Inter font
• Added light/dark mode switch to landing page
• Refined colors for better contrast
• New button variants for different purposes

Improved Context Awareness

Context Awareness

• Enhanced conversation context understanding
• Improved relevance of AI responses
• Better conversation coherence
• Maintains context across longer discussions

Message Categorization and Landing Page Updates

Message Categorization

• Added message type categorization
• Improved AI response relevance

Text Formatting

• Enhanced AI response formatting

Landing Page

• Replaced table with animated beam graph
• Updated sidebar background to match chat interface
• Fixed footer text alignment

Model Selection

• App now remembers model selection between landing page and chat

Cookie Settings and Chat Improvements

Cookie Consent

• All options enabled by default when opening settings
• Clearer organization of cookie preferences

Message Display

• Added gradient fade for out-of-bounds messages

Chat Management

• Fixed issue with editing multiple chat names simultaneously

PWA Improvements

• Improved icon handling for better browser compatibility
• Enhanced visual consistency

User Tips and Improved Search

Tips for Users

• Added tips section under chat input
• Auto-changes every 30 seconds with fade effect
• Provides usage suggestions

User Preferences

• Created dedicated preferences system
• Added preferred AI model setting
• Improved font type and Enter key behavior settings

Chat Search

• Created dedicated search page for past conversations
• Visual highlighting of recent chats
• Improved search organization
• Better navigation between search and chat

Chart Visualization and Trial Period

Chart Visualization

• Added chart visualization for data analysis
• Copy chart images to clipboard
• Download chart images

Response Formatting

• Improved text formatting in AI responses

Landing Page

• Enhanced visual design
• Better alignment and spacing
• Improved mobile responsiveness

Chat Experience

• Better conversation continuity
• Improved handling of complex discussions

Trial Period

• 14-day trial for new users

Enhanced AI Understanding and Mobile Fixes

AI Improvements

• Enhanced conversation understanding
• Improved prompt handling for text and code
• More accurate responses based on conversation history

Message Processing

• Improved message categorization
• Better request understanding

Mobile Experience

• Fixed zoom issues on iPhones

User Interface

• Now displays preferred AI model in chat interface

Feedback System and Keyboard Shortcuts

Feedback System

• Added feedback system in sidebar
• Five feedback options with detailed description
• Submit feedback directly from the app

Keyboard Shortcuts

• Configurable Enter and Shift+Enter behaviors
• Customizable keystroke actions
• Persistent preferences across sessions

PWA Experience

• Improved install button visibility
• Better styling matching the interface
• Clearer installable state indication

Pricing

• Updated pricing boxes for easier signup

Chat Landing Page and PWA Support

Chat Experience

• Added landing page with sample prompts
• Sample prompts create new chats automatically
• New chat button now leads to landing page

PWA Support

• App can now be installed on devices
• More native app-like experience
• Improved performance

Navigation

• Chats now have their own URLs for easier access
• Redirect to main chat page for non-existent chats
• Redesigned sidebar for better mobile and desktop use

Error Handling

• Clearer error messages
• Better recovery options

Subscription Plans

Subscription Plans

• Added subscription plans for premium features
• Secure payment processing
• Multiple subscription tiers
• Simple subscription management

Chat Editing and Theme Improvements

Chat Management

• Added inline chat title editing
• Edit directly from chat list

Code Blocks

• Theme-aware syntax highlighting
• Matches light/dark mode preference

Cookie System

• GDPR-compliant cookie consent
• Options for all, reject optional, or customize

Profile

• Added ability to change first name in settings

Mobile

• Added "new chat" icon to mobile top bar

Model Selection and Stop Button

Model Selection

• Choose which AI models appear in chat interface
• Customize available models in settings

Input Handling

• Added token limit for prompts
• Prevents issues with extremely large inputs

User Experience

• First name collection during signup
• More personalized experience

Stop Button

• Full functionality for stop button
• Stop incoming message streams
• Cancel messages (won't be saved)
• Get previous prompt back after canceling

Context Awareness

Context Awareness

• Added first version of context awareness
• Better understanding of conversation context
• Works when switching between models mid-conversation

UI Improvements and Support Options

Chat Experience

• Added skeleton loaders to chat and sidebar
• Fixed scroll issues in chat view

User Interface

• Added Support & Community section in sidebar
• Email and Discord support options

Performance

• Optimized chat deletion with instant feedback

Welcome

This is Octofy's changelog where we document new features and improvements. We're currently building our team and working on an MVP. Check back for updates.