Speaker Search & Editable Transcripts — 100% Offline

TLDR; Alter’s transcript game just leveled up! Now you can rename titles, change summaries, and hunt down specific speakers or keywords—all while keeping your data 100% local and offline. Search has also been extended to your conversations’ content, not just titles. ✨

Tools

• Quick Actions system is available to everyone
• New YouTube Transcript system with timestamps (slower but working on almost all videos)

Search & History

• Find past conversations by searching titles and content
• Find past transcripts by searching titles, summary, content, and speakers
• Transcripts’ title and summary can be edited

User Interface & Experience

• Removed follow button outside workspace
• Fixed reuse functionality - now properly gets message content when reusing for another action

Bug Fixes & Stability

• Fix a notch glitch on open
• Fixed error reporting during conversation and fallback mechanism
• New health checks on servers

Date: 2025-06-13

Local Tools Are Now FREE For Everyone!

TLDR; Alter makes all Local Tools completely FREE for everyone - 17 powerful automation features including Messages, Calendar, Maps, and Contacts with no subscription walls. Unlike other AI chatbots, Alter actually controls your Mac through natural commands.

New Tools & Integrations

• Web Search Tool - Added local web search with DuckDuckGo
• Email Integration - Added get unread emails tool
• System Notifications - Added Apple system notification tool
• Apple Maps Integration - Added Apple Maps tools for location services and directions
• iMessages Support - Added iMessages integration for messaging
• Contact Management - Added Contacts tools to manage contacts

Tool Management & Organization

• Reworked tool categories for better organization
• Fixed tools removal from search functionality
• Fixed Apple Script tool form file picker

Search & History

• Added search by content using summaries in history

User Interface & Experience

• Restored selected files in templates for actions
• Removed follow button outside workspace
• Fixed reuse functionality - now properly gets message content when reusing for another action

System & Performance

• Fixed crash with out of bounds error when more speakers than colors
• Minor performance improvements
• Fixed context pasteboard reading to force read when getting it

Bug Fixes & Stability

• Fixed errors to keep message content after error occurs
• Fixed MCP start/stop on connect/disconnect

Date: 2025-06-06

Easier Transcriptions Access and Performance Improvements

TLDR; Alter introduces advanced transcription features and significant performance enhancements. Easily access your previous transcriptions, customize detection thresholds for dictation and meetings, and enjoy better tool management with a new global toggle.

Transcription and Dictation

• Added a transcription button and a list for easy access to previous transcriptions
• Search transcriptions anywhere, just like contexts
• Customizable thresholds between dictation and meeting modes
• Fixed language detection when using local Whisper

User Interface

• Added a toggle to disable all tools in the tool window
• Fixed input clearing and cmd+z crash after displaying context details
• Fixed text selection when using global shortcuts and prompt templates

System and Performance

• Fixed crashes when pasting URLs in the notch input
• Improved performance by eliminating double automatic model selection
• Updated the mcp library to fix crashes related to incorrect use of continuations
• Attempted fix for crashes occurring when waking from sleep
• Fixed original source for the grammar correction function
• Used the correct value for clipboard when creating templates in actions
• Fixed crashes when deleting parameters in the editor

Date: 2025-05-30

Auto-Insert AI Replies & Custom Forms

TLDR; Alter introduces two workflow game-changers this week: the Insert Quick Action and Form Builder.

The Insert feature automatically places the last AI reply into your active app. A great example can be seen with in the video below where Sam shows how to enhance the “Correct Grammar” action:

Meanwhile, the new Visual Form Builder lets you create interactive forms with radio buttons, dropdowns, and text inputs that can be used into your prompts as variables. Yup, now you can easily customize the “Mail Draft Reply” or “Change Tone” action.

This update also delivers critical stability improvements, fixing multiple crash scenarios and improving reliability across the board—because having Alter crash frequently is not part of our plan.

User experience

• Add documentation and examples for using context placeholders in the action editor
• fix search prefixes in the notch for long text inputs
• Extend the paste text limit to the notch input to 4000 bytes instead of 1000

Automation

• Fix Apple Script crash
• Apple Script can be used as tool by LLMs
• Fix context placeholders in custom actions
• Add Stencil string utilities to allow prompts with programmatic evaluation of contexts

System & Performance

• Multiple fixes on the conversation to avoid hangs and crashes
• Fix text selection and youtube context fetching
• Fix chatgpt-4o-latest with tools (again)
• Extend timeout to 3 minutes to avoid timeout with slow responses from LLM providers
• Performance fix for the message presenter (conversation window)
• Performance fix for checking the audio device
• Fix shortcuts settings cleared bug on reboot
• Fix tool runner memory issues
• Fix insert/replace behavior when Alter has a real window opened (tools, settings or action editor)
• Fix random hangs when creating a new action
• Fix icon issues in context cards

Date: 2025-05-23

Connect WhatsApp & thousands of Apple Shortcuts

TLDR; Alter now connects with Apple Shortcuts and webhooks, letting you automate workflows with WhatsApp, Make, N8N, Relay, Zapier and thousands of other apps.

Set up global keyboard shortcuts to trigger AI-powered actions that execute without lifting a finger. The update also includes voice improvements and critical bug fixes.

User experience

• Disable auto-scrolling on new user reply
• Copy as plain text a message in the conversation window
• Avoid the user reply input to hide when removing a context card
• Fix scrapping from link paste by replacing Jina reader with our own local HTML to Markdown engine

Voice

• Add a warning when the system language is not auto-detectable by the dictation engine
• Allow transcription language to be forced
• Improve Settings UI for the voice tab

Automation

• Add description examples for webhook and Apple Scripts Quick Actions
• Add working Webhook to Quick Actions (also usable as tools for LLMs)
• New card detail for tool calls
• Tool call detail card can be closed with the escape key press

System & Performance

• Fix random crash when awaking from sleep
• Fix crash when creating a new action
• Fix the templating system in custom actions
• Exclude chatgpt-4o-latest from tool support
• Fix window not closing on conversation deletion
• Fix Youtube crash when some data was missing

Date: 2025-05-16

Hotfix for custom API keys

This update patches the beta21 release from May 9th, where Alter ceased to work with local models or custom providers.

⚠️ This update includes a database upgrade that may take up to 10 minutes. Please allow the process to complete.

TLDR; Alter can now talk to 2 000+ apps. Early preview for Yearly and Lifetime users. Also introducing Quick actions, for everyone.

Full changelog

• Introducing Apps: connect with 2000+ external services
• Introducing Local tools: Apple Calendar, Finder, Mail, Notes, Reminders, Safari
• Quick actions: create custom actions for acting on the last assistant reply (create Apple mail draft, reuse with another Alter action, send webhook)
• New output languages: Czech, Danish, Dutch, Finnish, Norwegian, Polish, Swedish, Turkish, Vietnamese
• Link your Alter account with your Discord account to get exclusive access to a private channel (Yearly and Lifetime holders)

Date: 2025-05-09

🔌 2 000 Apps, ⚡ Quick Actions

TLDR; Alter can now talk to 2 000+ apps. Early preview for Yearly and Lifetime users. Also introducing Quick actions, for everyone.

• Introducing Apps: connect with 2000+ external services
• Introducing Local tools: Apple Calendar, Finder, Mail, Notes, Reminders, Safari
• Quick actions: create custom actions for acting on the last assistant reply (create Apple mail draft, reuse with another Alter action, send webhook)
• New output languages: Czech, Danish, Dutch, Finnish, Norwegian, Polish, Swedish, Turkish, Vietnamese
• Link your Alter account with your Discord account to get exclusive access to a private channel (Yearly and Lifetime holders)

Date: 2025-05-02

Whisper local models are here!

TLDR; Despite our busy personal week, we’ve added Whisper models for offline dictation - giving you better voice recognition without needing an internet connection. It’s free for everyone!

What This Actually Means:

Even when our team was celebrating life events (Gaël turning 40 and Kévin welcoming baby Nathan!), we kept our 22-week release streak alive with something you’ve been asking for.

Now you’ve got three ways to dictate:

1. Apple’s built-in option (which we know was disappointing)
2. Our new Whisper models (try “small” for the sweet spot)
3. Cloud processing (still the premium choice)

Just hit the dictation tab, pick your model, and start talking. The best part? This upgrade is for everyone - no subscription required.

Because sometimes you need to work offline, and Apple’s dictation just wasn’t cutting it.

User experience

• Support for Orion+, LivreWolf, Opera GX, Thorium and Beam browsers
• Code block wrapping
• New output languages: Chinese, Indonesian, Italian, Japanese, Portuguese, Russian and Vietnamese

System & Performance

• Improved Dictation and LLMs calls by 500ms all regions
• Added a temporary fallback system between OpenAI and VertexAI providers

Bug Fixes

• Deleting an action with a global shortcut unbind the shortcut
• Force activate app on what input
• Trim filename when saving to markdown

Date: 2025-04-25

Global Shortcuts & Enhanced Editor

TLDR; Supercharge your workflow with custom global shortcuts for any Alter action and enjoy a completely redesigned editor that makes creating custom actions a breeze.

We’ve reworked the action editor te make it easier and more intuitive.

What This Means For Your Productivity:

Global shortcuts let you trigger any Alter action from anywhere on your Mac - no need to open the app first. The redesigned editor brings a cleaner interface that puts the focus on your content.

New Editor & Global Shortcuts

The star of this release is our completely redesigned editor. Combined with global shortcuts, you can now:

• Assign custom keyboard shortcuts to any Alter action
• Trigger actions from anywhere on your Mac
• Easily edit the system and user prompts of your user-defined actions

Microphone Enhancements

• Right-click to cancel recording - no more awkward attempts to stop unwanted recordings
• Improved handling of function keys during recording

System & Performance

• Fixed infinite loader when only whitespace remains after removing ’/’
• Improved notch behavior with better current app detection
• Optimized shortcuts management to prevent conflicts
• Better handling of text selection context

Bug Fixes

• Fixed translation issues in German UI
• Improved search functionality when clicking on “chips” in the notch input
• Enhanced keyboard navigation for radio parameters

Date: 2025-04-18

Community updates + Product Hunt launch

TLDR; Next Wednesday, we’re officially launching Alter on Product Hunt! 🎉

We need your help to make some noise and show the world what Alter can do!

Here’s how you can be a HUGE help:

1. Mark Your Calendar: Next Wednesday! 🗓️
2. Sign Up for Notification: Head to our Product Hunt ‘Notify Me’ page: https://www.producthunt.com/products/alter-4
3. Support on Launch Day: Go to Product Hunt, give us an upvote 👍, and leave a genuine comment about how Alter helps (or how you’re excited for it!).
4. Spread the Word: Tell your network, friends, colleagues – anyone who needs an AI assistant that actually saves time.

Let’s crush this launch together! We need everyone’s support to climb those PH rankings. 💪

Thanks for being awesome! 🙏

And check our sneak preview of Alter Automations that will let you connect with 2,000+ web services:

User experience

1. New settings in appearance to remember the last position of the conversation window
2. Do not reinitialize the Ask Anything text field on model selection by click
3. Add icons to the models list

System

1. New model: Gemini 2.5 Flash Preview (us-central)
2. New model: o3 and o4-mini
3. Orion Plus: now supported like other browsers
4. Fix a bug with setting a custom endpoint

1. Perfect speaker identification for all your meeting
2. This AI Feature Just Saved Me 25 Minutes
3. Create the perfect logo for free with AI
4. Adding Context to Alter Conversations
5. Master System Instructions for AI Assistants: A Simple Guide
6. How to chat with Youtube Videos
7. How to chat with apps (or the fantastic follow-mode)
8. How to use Workspaces
9. A 2 minutes demo
10. How to chat with Files and Folders from the Finder

Date: 2025-04-11

Automatic Meeting Recording & New Models

TLDR; Alter automatically detects when you’re in a meeting and offers to record it - no more scrambling to hit record or missing important conversations.

This is not the default. You can enable by going to your settings:

1. Open Alter Settings > Dictation tab
2. Toggle “Auto Record Meetings” to ON
3. That’s it! The next time you join a meeting, you’ll see a notification
4. Do nothing (or click the X) and recording begins automatically

Pro Tips for Meeting Masters

• All recordings are automatically transcribed with speaker identification
• Navigate recordings with interactive timestamps
• Ask Alter questions about your meeting content later
• Find all your transcripts in the notch context list or in ~/Library/Application Support/Alter/Transcripts

This feature is highly experimental so let us know how it works for you

• Join our Discord community to share your experiences and get support: https://discord.gg/gvCMmfBRWZ
• Contact us directly at [email protected] for any feedback or suggestions

Voice & Dictation

1. Add Facetime to the watch list for meeting detection
2. Show a message when speakers names can’t be edited (conversation)

User experience

1. New settings to disable Youtube and Meeting notifications
2. Auto scrolling improvements in conversation windows
3. Fixed broken paste behavior with the Correct Grammar action
4. Add a shortcut to delete a conversation right away (CMD+RET)

System

1. New model: Gemini 2.5 Pro Preview (us-central)
2. New model: Grok 3 and Grok 3 mini
3. Fixed max_tokens error when the value is not supported by the model

1. Perfect speaker identification for all your meeting
2. This AI Feature Just Saved Me 25 Minutes
3. Create the perfect logo for free with AI
4. Adding Context to Alter Conversations
5. Master System Instructions for AI Assistants: A Simple Guide
6. How to chat with Youtube Videos
7. How to chat with apps (or the fantastic follow-mode)
8. How to use Workspaces
9. A 2 minutes demo
10. How to chat with Files and Folders from the Finder

Date: 2025-04-04

Alter + Gemini Web: The Smartest Way to Search Without Breaking Your Flow

TLDR; We’ve redesigned the experience to bring you real-time information from across the web, with cited sources, directly within your Alter workflows.

Our newest addition, Gemini 2.0 Flash Web, makes searching the internet faster and more powerful than ever before!

It is now our search model per default when using cmd+enter and is part of our “Best” model pipeline.

We’ve upgraded the transcript experience with speaker name editing, giving you more control over how your conversations are organized and displayed.

What This Means For Your Productivity:

• Real-Time Information: Get the latest data from across the web without ever leaving Alter
• Contextual Awareness: Alter understands what you’re looking at and combines it with web search results - Cited Sources: Every answer includes links to original sources—easily verify information or dive deeper
• Advanced Search Capabilities: Filter by specific websites, time periods, or content types (like YouTube videos)

Here’s five prompts to try:

1. Sports Updates: “When is the next game of Arsenal?” - Get the latest fixture information with official sources.
2. Context-Aware Research: While on a player’s profile page, simply ask “How many games with his national team has this player?” - Alter combines your current window context with web search.
3. Filtered News: “Explain Trump tariffs using only French news sources” - Limit results to specific publications or regions.
4. Recent Content: “Show me YouTube videos about Mickey 17 published in the last 24 hours” - Filter by recency and content type.
5. Academic Research: “Give me the latest research on ODA control systems published last week from PDF sources only” - Perfect for finding scholarly articles and papers.

Pro tip:

For the most powerful results, try combining your current context with web search. For example, when on a website about a specific topic, Alter can combine that context with web search results for truly personalized answers.

Voice & Dictation

1. Fixed transcript text and audio download in conversation window
2. Rework Transcript context detail card
3. Fixed issues with meeting transcription when using Best model

User experience

1. Close conversation windows with CMD+W instead of ESC
2. Add cards for sources web citations from Perplexity and Gemini
3. Add arrows to scroll citations in the conversation view
4. Add Help menu to the notch with direct links to the documentation

System

1. New model: Gemini 2.0 flash with grounding search
2. Performance optimizations across the app

1. Perfect speaker identification for all your meeting
2. This AI Feature Just Saved Me 25 Minutes
3. Create the perfect logo for free with AI
4. Adding Context to Alter Conversations
5. Master System Instructions for AI Assistants: A Simple Guide
6. How to chat with Youtube Videos
7. How to chat with apps (or the fantastic follow-mode)
8. How to use Workspaces
9. A 2 minutes demo
10. How to chat with Files and Folders from the Finder

Date: 2025-03-28

Transcript Management Refined

TLDR; Name your speakers in transcripts for clearer conversations and enjoy smoother keyboard shortcuts.

We’ve upgraded the transcript experience with speaker name editing, giving you more control over how your conversations are organized and displayed.

What This Means For Your Productivity:

• Personalized Transcripts - Label speakers with their actual names for better readability
• Improved Organization - Easily identify who said what in multi-person recordings
• Cleaner Documentation - Create professional meeting notes with properly attributed quotes

The Problem We Solved

Transcripts with generic “Speaker 1” and “Speaker 2” labels made it difficult to follow conversations and share meeting notes. Now you can edit these labels to match the actual participants, making your transcripts immediately more valuable and easier to reference.

Voice & Dictation

1. Speaker name editing in transcripts
2. Improved transcript organization and readability

System

1. Fixed global shortcut issues where GUI APIs weren’t properly called
2. Improved notch behavior - now closes if already open when using keyboard shortcut
3. Performance optimizations across the app

2. Adding Context to Alter Conversations
3. Master System Instructions for AI Assistants: A Simple Guide
4. How to chat with Youtube Videos
5. How to chat with apps (or the fantastic follow-mode)
6. How to use Workspaces
7. A 2 minutes demo
8. How to chat with Files and Folders from the Finder

Date: 2025-03-21

The Best Model

TLDR; Stop Thinking About AI Models: The right model, every time, without you lifting a finger.

We’re excited to announce a major upgrade: Best Model Selection! Alter now intelligently chooses the optimal AI model for your prompts and Alter Actions, delivering faster and more reliable results.

What This Means For Your Productivity:

• Faster: Experience up to 50% speed improvement on many actions like Correct grammar.
• Handles Complex Tasks: No more context window worries! Alter automatically selects the best model for large documents, multiple folders, and search requests.
• Vision by Default: Image understanding is now seamless. You don’t need to think about model capabilities – Alter handles it.

The Problem We Solved

Stop wasting time manually selecting AI models. Alter’s Best Model Selection takes the guesswork out of the equation, so you can focus on what matters: your work.

See It In Action:

Choose “Alter > best” from the model select list or type “/best”

What’s nice is the capabilities to automatically choose the right model at each turn of the conversation. It can start with a web search, use a smarter model for analysis and a fast one to format all in the same conversation.

While we’ve optimized Best Model Selection globally, performance can sometimes vary based on your location. If you encounter any issues, please contact us at [email protected] or join our Discord community (link below). We’re committed to making it work perfectly for you!

→ Click here to join

Voice & Dictation

1. Fix auto-following transcript replay

Conversation & Visual Experience

1. Rework of Perplexity Sonar support
2. Embed Perplexity citations links in the reply
3. Add Orion browser support for Youtube notifications

System

1. Fix Reuse follow-up where the context was not added to the notch
2. Add the user email to the profile section
3. Detect IME composition for non-european scripts
4. Optimize dictation speed performance

1. Create the perfect logo for free with AI
2. Adding Context to Alter Conversations
3. Master System Instructions for AI Assistants: A Simple Guide
4. How to chat with Youtube Videos
5. How to chat with apps (or the fantastic follow-mode)
6. How to use Workspaces
7. A 2 minutes demo
8. How to chat with Files and Folders from the Finder

Date: 2025-03-14

Smart Pasting

TLDR; Smart Pasting is here to save you from the daily frustration of broken formatting when copying and pasting content.

It’s a game changer for anyone who uses Alter and needs to share content across different apps.

What This Means For Your Productivity:

• Automatic Format Detection - Alter now understands where you’re pasting and adapts the format instantly
• Zero Configuration Required - No settings to tweak, no keyboard shortcuts to learn
• Time Saver - What used to take minutes of reformatting now happens instantly

The Problem We Solved

We’ve all been there. You copy a perfectly formatted reply from Alter, paste it into an email, and… markdown chaos. Headers turn into hashtags, formatting disappears, and you waste precious minutes fixing it all.

Last week alone, I spent 25 minutes reformatting meeting reports across 5 different emails. That’s when we decided: “We need to fix this.”

See It In Action:

When you copy content from Alter and paste it:

• Into Obsidian → Clean Markdown preserved
• Into Apple Mail → Perfect Rich Text formatting
• Into Google Docs → Clean text with formatting intact, no markdown symbols

We’ve tested this across dozens of apps, and it just works. Copy, paste, done.

This is one of those features that seems small but saves you time every single day – the kind you don’t notice until it’s gone.

Doesn’t work with one of your app, make your request in our Discord → Click here to join

Voice & Dictation

1. Better support for spelled-out punctuations
2. Support for quick commands like “Create a todo list in markdown with the 3 following entries…”
3. Fix a bug where the model would hallucinate its result when using a quick command

Conversation & Visual Experience

1. New shortcuts to change the font-size of the conversation window (cmd+plus, cmd+minus)
2. Replaced pins with stars to favorite a conversation or model
3. Conversations can be renamed both in the chat window and in the notch
4. Saving a reply will save in a Markdown file

System

1. Workspaces: performance improvements + fix crash with large code repositories (4K+ files)
2. Onboarding check for connection issues
3. Fix crashes around fetching browser tabs context
4. Fix race condition in context management
5. New shortcut to rename highlighted conversation in the notch cmd+R

1. Create the perfect logo for free with AI
2. Adding Context to Alter Conversations
3. Master System Instructions for AI Assistants: A Simple Guide
4. How to chat with Youtube Videos
5. How to chat with apps (or the fantastic follow-mode)
6. How to use Workspaces
7. A 2 minutes demo
8. How to chat with Files and Folders from the Finder

Date: 2025-03-07

Unleash Your Creativity with Ideogram Models

Image creation is going next level with Ideogram new version V2a.

It can generate premium graphic designs with text and photo-realistic images in just a few seconds.

It is a game changer for logo and visual asset creations.

What this means for your productivity:

• Faster generation - Iterate quicker and come up with the right assets
• Exceptional typography handling-Easily change a text in a design while maintaining perfect typography style
• Versatility - More than logos, Ideogram V2a is great at doing high quality images and poster

At Alter, we believe AI should not only enhance your productivity but also unlock your creative potential. These new image generation capabilities represent a significant step toward that vision.

We can’t wait to see what you create. Share your designs in our Discord → Click here to join

Voice & Dictation

1. Fix recording errors happening on server starts

Visual Experience

1. Support for larger text input conversation windows
2. Faster model search

System

1. Favorite your conversations
2. Add support for kagi’s Orion browser
3. Add filter system for custom provider models
4. Support for Azure custom endpoint
5. Fix Workspaces spamming our servers

1. A Look at the Alter new interface
2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-02-28

Drag, Drop, Done: Now In All Conversation Windows

You can now drag & drop or copy-paste any files or images directly into your existing Alter conversations!

This is the contextual AI experience we’ve been building toward - where your assistant adapts to your workflow, not the other way around.

What this means for your productivity:

• Continuous thought streams - No more restarting conversations to add new information
• Seamless context building - Drop in research, images, or code as your work evolves
• True workflow integration - Your AI assistant that grows with your thinking process

At Alter, we believe AI should fade into the background of your work - powerful enough to transform your productivity, yet natural enough that you forget it’s there. This update is a significant step toward that vision.

We can’t wait to see how this transforms your workflow. Share your experience in our Discord → Click here to join

Voice & Dictation

1. Another set of performance improvements
2. Prevent keep recording notification from being shown at start

Visual Experience

1. New notch mode: Hide me!
2. Add line wrapping in markdown code blocks
3. Show a deletion confirmation dialog when deleting a transcript or a conversation
4. Prevent notch closing immediately when opening
5. Fix inconsistent appearance in the Settings
6. Show Alter version in the menubar

System

1. When opened, the entire notch works as a drop zone.
2. Clamp number of images generated to match models capabilities
3. Support for OpenAI o1 and o3 models
4. Fix Keynote slides generation
5. Support Mistral as a custom provider
6. Fix some AppSense crashes

1. A Look at the Alter new interface
2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-02-21

Now Alter works on your terms

We’ve added Profile - a simple way to tell Alter who you are and how you want to work together.

It’s pretty straightforward:

1. Tell Alter about yourself
2. Share what you’re working on
3. Set your preferred communication style

Why? Because context matters. When Alter knows you’re a developer who prefers concise technical answers, or a writer who wants creative brainstorming, it can skip the fluff and be more helpful.

Try it out and let us know what you think. We’re still fine-tuning this, so your feedback helps.

We’ll share some profile templates on the Discord community → Click here to join

Voice & Dictation

1. Improve dictation transcription (2x faster)
2. Improve dictation paste
3. Add a “Keep recording” notification after recording for more than 90 minutes

Visual Experience

1. The notch can now be hidden
2. New onboarding steps with profile templates and custom actions
3. Fix the floating widget stuttering

System

1. Add a toggle to enable/disable a custom provider
2. Fix performance issues when pasting large amounts of text in the Ask Anything field
3. Fix big images crashing Alter (heic format)
4. Fix models fetching issues
5. Fix saving history issues

2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-02-14

Perfect voice control with smart entities

TL;DR: Alter understands you meant Anthropic’s Claude, not your brother Claude

By understanding the context of the page, voice feels magical:

• Alter understands what you’re referring to, even with partial commands or strong accents (oui, even our French accent works!)
• Alter intelligently distinguishes between similar-sounding words
• Alter masters acronyms and technical terms, from IT jargon to RapGenius’ slang.

Watch our quick video guide to see these features in action:

Voice & Dictation

1. Another round of fixes on recording
2. Added start and stop sound for recording
3. Trim generated text to avoid unwanted new lines
4. Added a notification to avoid accidental long recordings

Visual Experience

1. New onboarding experience
2. Added a Top Center placement for the conversation window
3. Updated the appearance of the notch

System

1. Fixed memory leaks when tabbing with Chromium-based apps or browsers
2. Fixed other memory issues around webviews and conversations
3. Updated the embedded help screen to explain main features

1. A Look at the Alter new interface
2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-02-07

Every Copy is Now an AI Superquery

TL;DR: No more copy-paste juggling - your clipboard content goes straight to AI.

Universal clipboard support is here. Simply copy anything and paste it directly into Alter’s notch.

• Universal clipboard support (text, files, folders, images)
• Instant OCR for screenshots
• Smart context building from clipboard content

Watch our quick video guide to see these features in action:

Voice & Dictation

1. Fixes to avoid audio recordings crashes
2. Big improvement over dictation
3. Smart formatting based on the active app
4. Support for dictating emojis

Navigation & Controls

1. Delete conversations, workspaces and transcripts with backspace
2. Support for tab key in form
3. Clear context list after use

Visual Experience

1. Prevent input shifts in the conversation view
2. Artifacts can be closed

System

1. Fixed memory leaks
2. Better performance listing past conversations
3. Better multi-language support for built-in prompts
4. New fix for saving conversations
5. Avoid multiple Youtube notifications on Brave
6. Support for pasting images and files in the notch (cmd+v)

1. A Look at the Alter new interface
2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-01-31

Your Meetings Just Got a Memory Upgrade

TL;DR: Your feedback shaped a big Meeting Recording update - with time-stamps, speaker tags, one-click audio navigation and lots of bug fixes!

Special thanks to @gwd586, @nathanredblur, @Ankit, @JDub and @KataKlizym who helped shape this week release!

Watch our quick video guide to see these features in action:

1. Track meeting dynamics

• See speaking time distribution
• Analyze topic coverage

2. Quote with confidence

• Find and verify exact statements

3. Late Entry Hall of Fame

• Track dramatic “Sorry I’m late” entrances (just kidding… or are we? 😉)

Voice & Dictation

1. Major improvements in audio recording reliability
2. Implementation of new dictation system

Navigation & Controls

1. Enhanced keyboard navigation smoothness
2. Menubar entries now accessible from the notch
3. More intuitive escape key behavior implementation

Visual Experience

1. Selected model now displayed in the notch
2. Added hover highlights for floating menu entries
3. Fixed display issues for long user messages

Workspaces

1. Improved indexing speed
2. Extended support for programming languages
3. Enhanced think step intelligence

Mail & System Integration

1. Resolved Apple Mail permission issues
2. Fixed Apple Pay & AppStore hidden button behavior
3. Enhanced paste functionality reliability
   • Now working in Apple Mail again
4. Fixed conversation history saving issues
   • Resolved artifacts and follow-ups problems
5. Eliminated unwanted Youtube notifications
6. Added support for pasting images and files in the notch (cmd+v)

1. A Look at the Alter new interface
2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-01-24

Spotlight-style Search for AI Models

We just shipped something cool: a Spotlight-style search for AI models.

What’s new:

• Type to find any model instantly with the slash sign ”/” in the Ask Anything field
• Pin your favorites for quick access
• Switch between models with keyboard shortcuts

It’s that simple - just like how you’d search apps with Spotlight, but for AI.

Big thanks to @nathanredblur and @JDub who suggested this. Your feedback shapes what we build.

Pro tip: I personally pin:

• Claude 3.5 Sonnet for everyday tasks
• Local model when I need privacy
• Deepseek R1 for complex thinking
• Cerebras when speed matters

System Improvements

1. Deepseek R1 model 164k context window
2. Sonar & Sonar PRO as new online search models
3. Performance improvements (overall search system in multi-screen environment)
4. Sound recording fixes
5. Faster title generation when using Alter-provided models
6. Optimized image payloads
7. Improved performance of the transcript detailed card

1. Rework of artifacts cards

2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-01-17

Transform any text with voice commands

You can now select any text and use voice commands to transform it instantly.

Try these magic words while clicking and holding the record button (or keeping the fn key pressed):

• “Rewrite in friendly tone”
• “Create counter arguments”
• “Add my thoughts: …”

Also we improved the support for keyboard-first users in the context-list:

1. Basic Navigation

   • Use Arrow keys to move through items
   • Press Enter to open selected item
   • Press Return to remove selected item
   • Use Space to select multiple items

2. Advanced Commands

   • CMD+J: Create a workspace with selected items
   • CMD+K: Clear the entire context list

• Right-click any selection
• Choose “Add to Alter” from services menu
• Can be customized with keyboard shortcut through macOS services settings

New Features

1. Writing category: Fill in action
2. Vivaldi browser support
3. Transcript timestamps
4. Codestral model with 256k context window

System Improvements

1. Performance improvements
   • Notch opening optimization
   • Text typing in main field
2. Smart notifications fixes
3. Sound recording fixes

User Interface Enhancements

1. Ask Anything field expansion beyond 3 lines
2. Markdown code block copy button restored
3. Multi-screen focus fixes

Accessibility

1. MacOS increased contrast support

2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-01-10

Meet your meeting scribe

We are improving Alter 2 to make it one of the best software for handling your meetings. Here’s what’s new:

1. Work across any meeting software, browsers included
2. Nudge you to record your meetings for transcription
3. Create your meeting report with actionable steps

We’ve also added smart notifications that suggest actions based on what you’re doing:

• Record a transcription when Alter thinks you’re in a meeting
• Generate a meeting report when you stop a long audio record
• Summarize YouTube video when you watch a video

• Smart notifications: suggests actions based on what you’re doing
• New action: Meeting report
• New models: DeepSeek 3 and Llama 3.1 with 8B 128k
• The Ask Anything field does not search conversations anymore for performance

• Duplicated and missplaced microphone icon
• MP4 files are now supported for manual transcription
• Mail actions from the menu are working again
• Performance improvements

2. How to chat with Youtube Videos
3. How to chat with apps (or the fantastic follow-mode)
4. How to use Workspaces
5. A 2 minutes demo
6. How to chat with Files and Folders from the Finder

Date: 2025-01-03

Many fixes and improvements

• New model: Gemini 2.0 Flash experimental
• Adding images to the notch shows the OCR text, if any.
• Dropping PDF files works again
• Added Korean and Spanish as output languages
• Big fixes and improvements over follow-mode on folders and files
• Fix unavailable actions on text selections
• If the dock is fixed, the conversation window will not overlap with it
• Reworked the settings view
• Analyze Paper action moved to “Explain” category

Alter not only allows you to converse about the content of any window but also keeps the conversation aware of any changes that occur in that window!

Video:

Workspaces are a way to save knowledge and apps to follow for future sessions. By default, each elements in the workspace is followed. See “follow mode” above for more details.

Video:

Alter - Workspaces

Date: 2024-12-27

Drop audio and get transcripts

You can add audio files in the notch and generate a transcript.

Plus we fixed a last minute bug that was preventing the creation of workspaces.

1. Add an audio file to the notch
2. Open the context card
3. Click on the “Generate transcript” button
4. Wait for the transcript to be generated
5. Download or use the transcript as context

Date: 2024-12-20

Work With Intelligence

Happy to introduce the first personal AI assistant living in your notch!

This new version is a design shift built on the learning from Alter v1. It brings new features and drops some others.

What’s new and exciting:

• Voice support: push-to-command, dictation and meeting transcriptions
• Workspaces: turn documents and apps into knowledge for recurrent use
• Follow-mode: keep chatting while Alter keeps the context updated by watching apps, folder and files changes
• Notch: drag and drop many different kind of objects to build extensive contexts
• Fast navigation
• Optimized menu for text selections

Removing:

• knowledge bases: it’s replaced by Workspaces
• Mnemonic keyboard shortcuts: replaced by the notch fast search and will introduce global shortcuts in next version
• automatic context retrieval: for privacy reason, we favor optin context retrieval. For example we don’t fetch selected mail automatically. You can either drag and drop the email to get the full thread or add the app for screenshot + appsense retrieval.

Date: 2024-12-13