What's new?
Changelog
Roadmap
Getting Started 👋
Installation & First Steps
Basic Settings
Chatting with Alice
Assistants
Snippets
Remote Actions ✨
Remote Actions Basics
Google Calendar scenario
Quick scheduling links
Using existing automations
Better way to pass response to Alice
FAQs and help 💬
All frequently asked questions
Advanced
Custom Backend
Examples
Simple Snippets
Vision Snippets
Getting Started 👋
Return to all Docs

Installation & First Steps

Getting Started with Alice Installation

Hey, it's Alice! To kick off our chat, please follow the simple steps I've laid out below. It'll only take a few minutes tops.

  1. Download: Get the "Alice" app install file from https://heyalice.app
  2. Installation: Open the downloaded file. Move the app to the Applications folder (macOS) or follow the installer steps (Windows).
  3. First launch: Double-click the Alice app file to open it. You might receive a security prompt; follow the on-screen instructions to grant permission.

Onboarding

When you first launch the Alice app, I'll message you first, walking you through some basic settings. Only the first setting is mandatory; you can skip the rest if you'd like.

  • Activate account (required): Here, I'll ask for your email. Please provide the exact one used when placing the order. If this address is inactive, email Adam and Grzegorz at alice@techsistence.com for assistance. They'll help!

  • Your Avatar: It would be great to see you, even as an avatar. So for the next step, I'll ask you to share a photo to display with your messages. I'll save this image locally on your device.

  • Exploring the possibilities: At the final onboarding stage, I'll outline my capabilities, highlighting various settings and options to customize what I can do to suit your needs.

Tip: After entering your email, press Command+N (macOS) or CTRL+N (Windows) to open a new chat with me and skip the onboarding.

Instant Chat

After activating your account, you can immediately start chatting with me. However, some features will be limited, and you can only send up to 100 messages. To unlock the full potential and remove these restrictions, simply add an API key to one or more of the platforms listed below.

Connect your own API keys

Below, you'll find the key steps to connect your OpenAI, Anthropic, Perplexity, or Groq accounts. For each, you'll need to provide your API key to access your account. Rest assured, I'll securely store them on your device.

OpenAI

My most advanced capabilities currently stem from OpenAI's GPT model family. These span conversation (including audio) and sophisticated image recognition. Here's how to use them:

  1. Register an account on platform.openai.com, then link a card and purchase a pre-paid credit of at least $5. Full activation may take a few minutes to several hours. Note: An OpenAI API account differs from a ChatGPT subscription. OpenAI API uses a "pay-as-you-go" model, but you can manage costs by pre-purchasing credits.
  2. Generate an API key on your account at platform.openai.com. Click "Create new secret key", name it "Alice", and grant it all available permissions.
  3. Go to the "Alice" app and open settings. In the General tab, you'll find the "OpenAI API Key" field. Enter your key there.

After adding the key restart the application. From now on, all my skills will be available to you.

Anthropic

Besides OpenAI models, you can use models developed by Anthropic, like Claude 3 Opus or Claude 3 Haiku. The first one features highly developed skills comparable to GPT-4-turbo (and sometimes better), but at a much higher price. Conversely, Claude 3 Haiku, suitable for simpler tasks, is comparable to the GPT-3.5-turbo model but with a significantly lower price point. Here's how to use them:

  1. Register an account on console.anthropic.com, then link a card and purchase a pre-paid credit of at least $20. Full activation may take a few minutes to several hours. Note: Anthropic API account differs from claude.ai service. Anthropic API uses a "pay-as-you-go" model, but you can manage costs by pre-purchasing credits.
  2. Generate an API key on your account at console.anthropic.com. Click "Create Key" and name it "Alice".
  3. Go to the "Alice" app and open your settings. Scroll down if needed. Below, you'll find the "Anthropic API key" input where you can paste your key.

From now on, the list of available models will include Claude 3 Opus, Claude 3 Sonnet and Claude 3 Haiku. In the future, the model name may change or be completely replaced by other models. This will require manual changes to the Snippets that use this model.

Groq

Groq is a service that enables work with selected open-source models. Currently, it features Llama 3 70B from Meta. Its standout feature is the inference speed (response generation), which can reach up to 900 tokens per second! Moreover, Groq is currently free but has a very aggressive rate limit restricting the use of models. Additionally, it can only be used for non-commercial purposes.

Important: Open-source models like Llama 3 have significantly lower capabilities compared to OpenAI or Anthropic models. However, this model will successfully handle less demanding tasks.

  1. Create a free account at console.groq.com.
  2. Go to the "API Keys" tab and generate a new key and name it "Alice".
  3. Go to the "Alice" app and open your settings. Scroll down if needed. Below, you'll find the "Groq API key" input where you can paste your key.

From now on, the list of available models will include llama3-70b-8192. In the future, the model name may change or be completely replaced by other models. This will require manual changes to the Snippets that use this model.