Configuration

Customize Agentic Browser to fit your workflow.

LLM Settings

Selecting a Provider

Agentic Browser supports multiple AI providers:

1. Open Settings in Agentic Browser
2. Go to AI Provider
3. Choose your preferred provider:

API Key Configuration

For cloud providers, enter your API key:

1. Go to Settings → API Keys
2. Enter your key for the selected provider
3. Click Save

Keys are encrypted and stored securely.

Model Selection

Choose the model for each agent:

Orchestrator: GPT-4o (best reasoning)
Web Surfer: GPT-4o-mini (faster, cost-effective)
Coder: Claude 3.5 Sonnet (strong coding)

Using Local Models

For Ollama users:

1. Ensure Ollama is running locally
2. Select Ollama as provider
3. Choose from available local models
4. No API key required

Browser Settings

Viewport Size

Configure browser dimensions:

Custom sizes also available.

User Agent

Set the browser user agent:

• Default: Standard Chrome user agent
• Custom: Specify your own string

Timeout Settings

Configure wait times:

Screenshots

Configure automatic screenshots:

• On action: Screenshot after each action
• On error: Screenshot when errors occur
• Manual only: Only when requested
• Disabled: No automatic screenshots

Workspace Settings

File Storage

Configure where files are saved:

Default: /workspace/browser-agent/
Custom: /workspace/your-folder/

Download Handling

Set download behavior:

• Auto-save: Download to workspace automatically
• Prompt: Ask before each download
• Block: Don’t download files

File Organization

Organize downloads by:

• Date: /downloads/2024-01-15/
• Task: /downloads/task-name/
• Flat: All files in one folder

Security Settings

Trust Levels

Configure approval requirements:

Strict Mode:

• All form inputs require approval
• All button clicks require approval
• Maximum security

Standard Mode (Default):

• Sensitive fields require approval
• Regular navigation is automatic
• Balanced security

Relaxed Mode:

• Only auth/financial actions pause
• Faster automation
• For trusted internal tools

Domain Management

Trusted Domains: Add domains where actions proceed automatically:

mycompany.com
internal-app.mycompany.com

Blocked Domains: Prevent navigation to specific sites:

malicious-site.com
competitor-internal.com

Action Filters

Block specific action types:

• Block all downloads
• Block form submissions
• Block external navigation
• Block script execution

Session Settings

Session Persistence

Control session handling:

• Fresh each time: New browser for each task
• Persist: Keep cookies and storage
• Restore: Start from saved state

Cookie Management

Handle cookies:

• Accept all: Accept cookie banners
• Reject all: Reject when possible
• Prompt: Ask for each site

Storage

Manage browser storage:

• Clear on task end: Wipe local storage
• Persist: Keep across tasks
• Manual: You control clearing

Performance Settings

Concurrency

Control parallel operations:

Resource Loading

Control what loads:

• Full: Load all resources (images, scripts)
• Lite: Skip images for speed
• Minimal: Essential only

Caching

Browser cache behavior:

• Enabled: Use normal caching
• Disabled: Fresh requests always
• Selective: Cache static only

Notification Settings

Desktop Notifications

Get notified when:

• Task completes
• Approval needed
• Error occurs
• Long-running task finishes

Sound Alerts

Enable audio notifications:

• Approval needed sound
• Task complete sound
• Error alert sound

Exporting Settings

Backup Configuration

Export your settings:

1. Go to Settings → Export
2. Choose what to include:
   • Browser settings
   • Security settings
   • Trusted domains
3. Download as JSON

Import Configuration

Restore or share settings:

1. Go to Settings → Import
2. Upload your config JSON
3. Select what to apply
4. Confirm import

Reset to Defaults

Return to default settings:

1. Go to Settings → Reset
2. Confirm reset
3. All settings return to defaults

API keys are preserved during reset.