Basics
Alfred is a personal AI assistant you can use through the Dashboard and connected platforms like Discord. Alfred comes with modules that extend his abilities, from browsing the internet, creating background tasks, to generating images or 3d models.
There are a handful of ideas that everything else sits on top of. Once these make sense, the rest of the docs are just detail.
Platforms
Section titled “Platforms”Platforms are places where you interact with Alfred.
- The Dashboard is the main web app for chat and settings.
- Discord lets you use Alfred around Discord with server context when enabled.
- GitHub lets Alfred work on your repositories.
It’s the same Alfred on all of them, with the same memory and the same settings, as long as the accounts are connected.
Models
Section titled “Models”Models are the AI engines Alfred can use. Different models have different strengths, speed, cost, input support, and plan availability.
Your selected model is a preference, not always a guarantee. If you’re near a usage limit, Alfred may temporarily use a more cost-effective model to keep your allowance available for longer.
See Models.
Modules
Section titled “Modules”Modules are Alfred’s abilities. They decide what Alfred is able to do for you, such as searching the web, using memory, setting reminders, running code, generating images, or working with files.
You can manage modules from the Dashboard. Turning off the ones you don’t use makes conversations cheaper and usually better. See Modules.
Cards are Alfred’s files. Uploads, generated outputs, saved notes and working documents all live as cards, which is how Alfred handles large amounts of text without pasting it into the conversation.
See Cards.
Brain is Alfred’s memory. After conversations he reflects on what happened and stores what’s worth keeping, so over time he builds up a picture of you and of himself.
See Brain.
Automation
Section titled “Automation”Automation is Alfred doing things when you’re not there. A reflex is code that runs when something happens, and a background agent is an AI worker that a reflex can wake when a decision actually needs thinking.
See Automation.
Usage is based on the real cost of model and tool activity. A short answer on a cheap model costs less than a long answer that uses a stronger model, reads files, searches websites, runs code, or generates media.
See Usage.