User Experience & Privacy
- Native macOS app (not Electron)
- Light and dark themes based on your system config
- Fast & Lag-Free
- Sandboxed (it doesn’t have access to anything on your system, unlike most other apps currently available)
- It only interacts with OpenAI servers, we do not have access to a single piece of your data
- All data is saved locally on your machine
- OpenAI does not save (or use) conversations through an API to train models (again privacy win)
- Every feature in the app has a keyboard shortcut to give you a “one-click” access
All ChatGPT Models
Work with all OpenAI GPT models. Switch models with a single click (⌘1)
Model Configuration & Profiles
Create AI profiles and configure:
- system prompt
- max tokens
- history size (how many messages to include in context for cost control)
- Switch between profiles with a single click (⌘2) or type ‘@’ into the chat.
This allows you to save a lot of time by pre-configuring the profiles for any use case you need: Translator, Developer, Proofreader, etc.
Beam can render all markdown tags and tables.
Templates (Prompt Library)
Templates are prompts with variables. You can save your prompts and mark the places that should be filled in. The next time you need to use the prompt, you only need to fill in the blanks (⌘3) or type ‘/’ into the chat.
- Use [clipboard] in templates to automatically pre-fill the template with contents of your clipboard
- Use [variable_name] inside the template text to create variables
Interact with other apps and documents
- Insert ChatGPT responses to any input on your system
- Select text anywhere in your system and chat about it
- Open PDF files on your system and chat about them
Macros are quick prompts you can run with your text selection. For example, select the text you need to summarize, press a hotkey, type “sum,” and it will ask GPT model to summarize the selected text for you.
Extensions are like “mini-prompts” with pre-defined values that you can use (and combine) to guide the model in specific directions.
Beam ships with a few sample extensions: tone, language, and format. For example, you can type # into chat, select Format, select Table, and the model will be instructed to reply using the table format. The prompt is based on the “template” you define when creating an extension.
- Toggle extensions using # in chat or ⌘E keyboard shortcut.
- It is also possible to add extensions to the templates using [#extension-name] format when creating a template.
- Create and create combine various prompts - creativity is the limit!
- Saved locally on your machine
- All conversations are saved automatically
- Quick access with a single hotkey press
- Fulltext search
- Keep the app in the dock or make it a menu bar application only
- Opt-out of auto history
- Adjust context prompt
- Change font size
- Adjust global shortcuts
- Use a floating chat window (like the launcher) to interface with GPT models without distractions quickly.
- Create and manage permanent chat windows
Command Palette (⌘K)
- All actions in the application are always available in the command palette.
- Every action has a keyboard shortcut to speed up your workflow
Global keyboard shortcuts (configurable)
- Toggle floating chat panel, defaults to ^␣
- Toggle the floating chat panel with context, i.e., select some text anywhere on your system and chat about it. Defaults to ^⇧␣
- Toggle the floating panel and use the macro for context selection, defaults to ^⇧M
Local Keyboard shortcuts & actions
- Insert last message ⌘⏎
- Copy last message ^C
- Insert the last code box ⌘⇧⏎
- Pin window ⌘P
- Change GPT model ⌘1
- Change AI Profile ⌘2 or type
@into the chat
- Use template ⌘3 or type
/into the chat
- Use extension ⌘E or type
#into the chat
- Open PDF file ⌘O
- Copy last code block ^⇧C
- Open conversation from history ⌘H
- Clear conversation ⌘R
- Save conversation ⌘S
- Open main app window ⌘B
- Regenerate the last message ⌘W