Interface Guide

After deployment, open the interface shown on your dashboard. This guide covers ComfyUI workspaces and OpenWebUI chat. The interface and manual setup steps depend on the selected configuration.

ComfyUI for Image & Video

ComfyUI represents a generation workflow as connected nodes. You can inspect model files, input images, prompts, and output settings before queuing a run.

Getting Started with ComfyUI

Access your deployment - Wait for the interface to be ready, then use its link on your dashboard
Load the default workflow - Read the template’s setup notes. Use the supplied graph if one is loaded, or import the named workflow JSON and select its inputs
Enter your prompt - Find the text input node (usually labeled "CLIP Text Encode")
Click Queue Prompt - Use the workspace’s queue/run control after checking all required inputs
View your result - Output appears in the preview node when complete

Key ComfyUI Features

Node-Based Workflows

Connect nodes to build custom generation pipelines. Right-click to add new nodes.

Workflow Import/Export

Save workflows as JSON files. Share or back up your custom setups.

Custom Nodes

Supported workflow deployments can install recognized custom nodes. Unknown, private, or incompatible nodes may need correction before the graph can run.

Queue Management

Queue multiple generations. Monitor progress in the queue panel.

Common ComfyUI Tips

Seed Control: Set a specific seed and save the graph and model versions to compare runs. A seed alone does not ensure identical output across configurations
Steps: Use the workflow’s documented setting. Distilled models can require a specific small step count; increasing it can make the result worse
CFG Scale: Start with the recipe’s value. The useful range depends on the model and sampler
Save Workflows: Use the menu to save your workflow before making changes

OpenWebUI for Text Models

OpenWebUI provides a ChatGPT-like interface for text generation models. Choose an installed model and start a conversation; available controls depend on that workspace’s configuration.

Getting Started with OpenWebUI

Access your deployment - Click "Open UI" on your dashboard
Create an account - If the workspace asks you to register, that account belongs to the deployed interface and is separate from your ModelPilot account
Start chatting - Type your message and press Enter or click Send
Create new chats - Click "New Chat" to start fresh conversations

Key OpenWebUI Features

Conversation History

Use the sidebar to find saved chats in that workspace. Export important conversations before deleting its storage.

Model Selection

Switch between available models using the dropdown. Different models have different capabilities.

System Prompts

Customize model behavior with system prompts. Set personality or specific instructions.

File Upload

Upload documents for the model to analyze. Supported formats include PDF, TXT, and more.

OpenWebUI Tips

Temperature: Lower = more focused, higher = more creative (adjust in settings)
Context Length: Models remember recent messages. Long chats may need new threads
Keyboard Shortcuts: Shift+Enter for new lines, Enter to send
Export: Export conversations as markdown or JSON from the chat menu

API Access

Use the API address and access instructions shown for your deployment. Native workspace APIs and ModelPilot’s generation APIs have different request formats:

Text chat through ModelPilot

/api/v1/chat/completions

Send your ModelPilot API key and an owned, deployed text model ID. An arbitrary model name does not create a new deployment.

Image Models (ComfyUI API)

Use the workspace API address shown on your dashboard

REST API for queuing workflows and retrieving results.

See our API Documentation for detailed integration guides.

Ready to try?

Try our free demo when available, or review a workspace configuration in the Model Catalog.