User Interface Organization

The PromptFx User Interface is divided into:

• A navigation bar at the top, allowing you to move forward/backward between different views, enter a "full screen" mode, and manage the default models used for text completion and embeddings.
• A navigation bar at the left, where you can navigate to different views.
• A main view, which is generally divided into inputs and outputs.
• A parameters panel at the right, where you can adjust model settings, prompt settings, etc. for the view.
• A "Run" button and status bar, for executing tasks using remote APIs.

PromptFx’s interface is modular: most tabs and views (including custom prompt apps) can be extended or overridden at runtime. Place your own views.yaml in config/ to add or replace UI panels. See Configuring Views for details.

Views

PromptFx has views for testing AI/ML models and for a number of basic applications. These views are organized into tabs on the left side of the UI:

• API Tab - Access to model playgrounds and API-based tools (completions, chat, embeddings, images, audio, etc.).
• Prompts - Tools for prompt library management, template editing, batch prompt scripting, and prompt validation.
• Text - Natural language processing tasks, including summarization, translation, entity extraction, sentiment analysis, and more.
• Documents - Document Q&A, document search, clustering, text manager, and PDF/snippet preview tools.
• Multimodal - Image and audio features, such as speech-to-text, text-to-speech, image generation, and description (Vision/Audio tabs).
• MCP - Support for Model Context Protocol (MCP) servers and prompts/resources/tools. (See https://modelcontextprotocol.io/.)
• Agents - Agentic workflows, tool chaining, and integrations with external APIs or agent plugins.
• Settings - Application and session configuration, system info, and runtime diagnostics.
• Custom - Uncategorized and user-configured views.