Web UI Guide

Browser-based research interface for non-engineers.

Overview

The Web UI provides a user-friendly browser-based interface for Zorora, designed for users who prefer a graphical interface over the command line. It offers the same deep research capabilities as the Terminal REPL with an intuitive visual interface.

Zorora Web UI - Browser-based research interface

Starting the Web UI

Method 1: Using zorora Command

zorora web

Method 2: Using Python Directly

python web_main.py

Access the Interface

Open your browser and navigate to:

http://localhost:5000

Features

Research Query Interface

Enter Research Question:

1. Type your research question in the search box
2. Select depth level:
   • Quick - Initial sources only (depth=1, ~25-35s)
   • Balanced - + Citation following (depth=2, ~35-50s) - Coming soon
   • Thorough - + Multi-hop citations (depth=3, ~50-70s) - Coming soon
3. Click “Start Research”

View Results:

• Synthesis with citations
• Source list with credibility scores
• Source URLs for verification
• Research metadata (timestamp, depth, source count)

Settings Modal

Access Settings:

• Click the ⚙️ gear icon in the top-right corner

Configure Models:

• Select models for each tool:
  • orchestrator
  • codestral
  • reasoning
  • search
  • intent_detector
  • vision
  • image_generation

Configure Endpoints:

• Select from:
  • Local (LM Studio)
  • HuggingFace Endpoints
  • OpenAI Endpoints
  • Anthropic Endpoints

Manage API Keys:

• Configure API keys for all providers
• Keys are masked for security
• Use show/hide toggle to view keys

Add/Edit Endpoints:

• Click “Add New Endpoint”
• Select provider
• Fill in endpoint details
• Click “Save”

Save Configuration:

• Click “Save” button
• Changes take effect after server restart
• Config file backup created automatically

Research Results Display

Synthesis Section

• Comprehensive answer with citations
• Inline citations ([Newsroom], [Web], [Academic])
• Confidence levels
• Key findings highlighted

Sources Section

• List of all sources used
• Credibility scores for each source
• Source type (academic, web, newsroom)
• Direct links to source URLs
• Credibility categories (high, medium, low)

Research Metadata

• Research ID
• Query text
• Depth level used
• Total sources found
• Completion timestamp

Settings Configuration

Model Selection

Orchestrator:

• Main routing and orchestration model
• Recommended: Local 4B model (fast, efficient)

Codestral:

• Code generation specialist
• Recommended: HF 32B model (high quality)

Reasoning:

• Research synthesis model
• Recommended: Local or remote based on quality needs

Other Models:

• Search, intent_detector, vision, image_generation
• Configure based on your needs

Endpoint Management

Local (LM Studio):

• Fast, no API costs
• Requires LM Studio running
• Privacy-preserving

HuggingFace:

• Remote inference endpoints
• Requires HF token
• Supports custom models

OpenAI:

• OpenAI API endpoints
• Requires OpenAI API key
• High quality models

Anthropic:

• Anthropic API endpoints
• Requires Anthropic API key
• Claude models

API Key Management

Security Features:

• Keys are masked by default
• Show/hide toggle for viewing
• Secure storage in config file
• Automatic backup before writes

Best Practices

Research Queries

• Be specific and clear
• Include context when needed
• Use depth levels appropriately:
  • Quick for fast answers
  • Balanced for thorough research
  • Thorough for comprehensive analysis

Configuration

• Start with local models (fast, free)
• Add remote models for quality when needed
• Use hybrid approach (local orchestrator + remote specialists)
• Test configuration before production use

Settings Management

• Backup config before major changes
• Test endpoints before saving
• Verify API keys are correct
• Check model availability

Troubleshooting

Web UI Not Starting

Problem: Server fails to start

Solution:

• Check Flask is installed: pip install flask
• Verify port 5000 is available: lsof -i :5000
• Check for errors in terminal output

Settings Not Saving

Problem: Configuration changes not persisting

Solution:

• Check file permissions on config.py
• Verify config file location
• Check for syntax errors
• Restart server after saving

Research Not Working

Problem: Research queries fail

Solution:

• Check model endpoints are configured
• Verify API keys are correct
• Check network connectivity
• Review browser console for errors

Endpoint Errors

Problem: Endpoint connection fails

Solution:

• Verify endpoint URL (for HF endpoints)
• Check API keys are configured
• Ensure endpoints are enabled
• Check API rate limits

See Also

• Terminal REPL - Command-line interface
• Configuration - Configuration guide
• Research Workflow - Deep research capabilities
• API Reference - Programmatic access