AI setup in Alpha Anywhere

Description

Configure IDE and Project providers so Alpha Anywhere can use AI locally or via hosted services. This walkthrough covers the AI menu, IDE efforts, provider selection, and project deployment providers.

Open The AI Menu

In the Web Project Control Panel toolbar, click the AI option to open the AI menu.

• AI IDE settings

  Profiles the IDE uses to help with development tasks.

• AI Project Settings

  Providers that ship with your application for deployment.

AI Menu In Web Project Panel

IDE Settings

Define IDE profiles and efforts for development-time AI assistance.

1. Open IDE Settings

   Choose AI IDE settings from the AI menu.

   

   AI IDE Settings Menu Item

2. Review Profiles

   The IDE settings shows a list of profiles so you can use different providers and models for different AI tasks.

   

   IDE Profiles

3. Define Default Effort

   Set up the fallback profile. The first level for IDE is effort: define a basic effort; optionally add better and best efforts for higher-quality models that may be slower or cost more.

   

   Effort Levels

4. Pick Provider And Model

   For the basic effort, select a provider, choose a model, and save any required API keys and endpoints.

   

   New Provider Setup

5. Example: Local Llama-CPP

   This example uses a local model running on llama-cpp on the local network.

   

   Select Llama-CPP Provider
   

   Specify Endpoint

6. Model Availability

   Llama-cpp servers often expose a small set of models; providers like OpenAI typically list many more.

   

   Available Models

Project Providers (Deployment)

Define providers your deployed application will use.

1. Open AI Project Settings

   Pick AI Project Settings from the AI menu to configure deployment providers.

   

   AI Project Settings Menu Item

2. Start With Empty List

   Project providers are developer-named; the list is empty until you create one.

   

   Empty Project Providers

3. Create A Provider Profile

   Select Create and enter a name, for example LocalDeployment.

   

   Name The Provider

4. Select Local LLM And Model

   Define a local LLM running on your machine; choose from available models (e.g., glm-flash in this example).

   

   Choose Model (Ollama)

5. Commit Provider

   After committing, the AI provider appears and can be published with your project. AI providers are a specialized case of NamedProviders and show in Web Project properties similar to storage providers.

   

   Project Provider Added