Custom Prompts

Create reusable system prompts, quick prompts, and copilot actions

Page Assist offers a powerful prompt system that allows you to create reusable prompts for various use cases. Whether you need consistent AI behavior, quick templates, or context menu actions, the prompt system has you covered.

Types of Prompts

Page Assist supports three distinct types of prompts:

Custom Prompts

System and quick prompts for use within the chat interface

Copilot Prompts

Built-in context menu actions for common tasks

Custom Copilot

Create your own context menu actions

This page focuses on Custom Prompts and Copilot Prompts that you can use within Page Assist’s chat interface and context menu.

Custom Prompts Overview

Custom prompts are templates you create and use within Page Assist’s chat interface. They come in two varieties:

System Prompts

Purpose: Define AI behavior and context for conversations

How they work:

• Sent as system role messages to the AI
• Set the personality, expertise, and constraints
• Applied at the start of conversations
• Persist throughout the chat session

Use cases:

• “You are a Python expert specializing in data science”
• “You are a creative writing assistant focused on sci-fi”
• “You are a technical documentation reviewer”

Quick Prompts

Purpose: Reusable message templates with variables

How they work:

• Inserted into your message input
• Support variables in {curly braces}
• Variables are highlighted for easy replacement
• Can be sent directly or edited before sending

Use cases:

• “Explain in simple terms”
• “Review this code for bugs”
• “Translate the following to “

Creating Custom Prompts

1. Access Prompt Management

   1. Open Page Assist Web UI
   2. Go to Settings
   3. Navigate to “Manage Prompts”
   4. Click the “Custom” tab
2. Choose Prompt Type

   Select either:

   • System Prompts - For AI behavior configuration
   • Quick Prompts - For message templates
3. Create Your Prompt

   Click “Add New Prompt” and fill in:

   • Title: Descriptive name for your prompt
   • Content: The actual prompt text
   • Variables: Use {variable_name} for placeholders (Quick Prompts only)
4. Save and Use

   • Click “Save”
   • Your prompt appears in the prompt selector
   • Select it when starting a chat or composing a message

System Prompts

System prompts define how the AI behaves throughout a conversation.

Creating System Prompts

Example 1: Code Review Assistant

You are an expert code reviewer with 15 years of experience.
Focus on:
- Security vulnerabilities
- Performance optimizations  
- Code maintainability
- Best practices

Provide specific, actionable feedback with examples.
Current date: {current_date_time}

Example 2: Academic Writing Helper

You are an academic writing assistant specializing in research papers.
Help with:
- Thesis development
- Argument structure
- Citation formatting
- Clear, formal language

Always cite sources and maintain academic tone.

Example 3: Customer Support Bot

You are a friendly customer support representative.
Guidelines:
- Be empathetic and patient
- Provide clear, step-by-step solutions
- Escalate complex issues appropriately
- End responses with "Is there anything else I can help with?"

Current time: {current_date_time}

System Prompt Variables

System prompts support time-based variables that auto-populate:

Variable	Description	Example
{current_date_time}	Full date and time	”March 3, 2026, 10:30 AM”
{current_year}	Current year	”2026”
{current_month}	Month number (0-11)	“2” (March)
{current_day}	Day of month	”3”
{current_hour}	Hour (0-23)	“10”
{current_minute}	Minute (0-59)	“30”

These variables are automatically replaced with current values when the conversation starts.

Using System Prompts

Web UI

1. Start a new chat
2. Click the system prompt selector dropdown
3. Choose your custom system prompt
4. Begin conversation with that context

Sidebar

1. Open sidebar
2. Click settings icon
3. Set default system prompt for copilot
4. Or select per-conversation in prompt dropdown

Quick Prompts

Quick prompts are reusable message templates with variable placeholders.

Creating Quick Prompts

Example 1: Code Explanation

Explain the following {programming_language} code in detail:

{code}

Include:
- What it does
- How it works
- Potential improvements

Example 2: Translation Request

Translate the following text to {target_language}:

{text}

Maintain the original tone and style.

Example 3: Content Summarization

Summarize the following {content_type} in {length} words:

{content}

Focus on the key points and main takeaways.

Example 4: Bug Report Analysis

Analyze this bug report for {project_name}:

Issue: {issue_description}
Steps to reproduce: {steps}
Expected: {expected_behavior}
Actual: {actual_behavior}

Provide:
- Root cause analysis
- Suggested fix
- Prevention strategies

Variable Syntax

Variables in quick prompts:

• Enclosed in {curly braces}
• Automatically highlighted in the UI
• Clicking highlights them for easy editing
• Use descriptive names: {user_input}, {target_language}, {code_snippet}

Use descriptive variable names to make templates self-documenting: {target_language} is better than {lang}.

Using Quick Prompts

1. Access Quick Prompts

   Click the “Quick Prompts” button in the input area (lightning bolt icon)

2. Select Template

   Choose your quick prompt from the dropdown list

3. Fill Variables

   • Variables are automatically highlighted
   • Click each variable and replace with actual content
   • Or edit the entire message as needed
4. Send Message

   Submit when ready - variables don’t need to be replaced if you prefer to keep them

Advanced Prompt Techniques

Chaining Prompts

Combine system and quick prompts for powerful workflows:

System Prompt: “You are a technical writer specializing in API documentation”

Quick Prompt: “Document this API endpoint: ”

Result: Technical documentation in consistent style

Contextual Prompts

Create prompts that work with Page Assist features:

With Knowledge Base:

Using my uploaded documentation about {topic}, answer:
{question}

Provide specific page references.

With Internet Search:

Search for recent developments in {field} and compare with {my_understanding}

With Vision:

Analyze this {image_type} image and identify:
- {aspect_1}
- {aspect_2}
- {aspect_3}

Role-Based Prompts

Create prompts for different personas:

Developer Persona

You are a senior {language} developer at a {company_type}.

Expertise:
- {framework_1}
- {framework_2}
- Best practices and design patterns

Provide production-ready code with explanations.

Teacher Persona

You are a patient teacher explaining {subject} to {audience_level} students.

Approach:
- Use simple language and analogies
- Provide examples
- Check for understanding
- Build on fundamentals

Current lesson: {topic}

Analyst Persona

You are a {industry} analyst with expertise in {specialization}.

Focus on:
- Data-driven insights
- Market trends
- Risk assessment
- Strategic recommendations

Analyze: {subject}

Prompt Library Examples

Programming & Development

Code Review

Type: System Prompt

You are a meticulous code reviewer focusing on:
- Security best practices
- Performance optimization
- Code readability
- Test coverage

Provide constructive feedback with specific examples.

Debug Helper

Type: Quick Prompt

Help me debug this {language} error:

Error: {error_message}

Code context:
{code}

Explain the cause and provide a fix.

Documentation Generator

Type: Quick Prompt

Generate {doc_type} documentation for:

{code_or_api}

Include:
- Description
- Parameters/Arguments
- Return values
- Examples
- Edge cases

Writing & Content

Content Editor

Type: System Prompt

You are a professional editor specializing in {content_type}.

Improve:
- Clarity and flow
- Grammar and style
- Engagement and impact
- SEO when relevant

Maintain the author's voice.

Blog Post Outline

Type: Quick Prompt

Create a blog post outline about {topic} for {target_audience}.

Requirements:
- {word_count} words
- {tone} tone
- Include {key_points}

Format with H2s and H3s.

Research & Analysis

Research Assistant

Type: System Prompt

You are a research assistant specializing in {field}.

Help with:
- Literature review
- Data analysis
- Citation management
- Hypothesis development

Always cite sources. Current date: {current_date_time}

Data Analysis

Type: Quick Prompt

Analyze this {data_type} data:

{data}

Provide:
- Key insights
- Trends and patterns
- Anomalies
- Recommendations

Use {analysis_method} approach.

Managing Prompts

Organizing Prompts

Naming Convention: Use descriptive, searchable names like “Code Review - Python” instead of “Review1”.

Categories: Prefix prompts with categories: “DEV: ”, “WRITE: ”, “RESEARCH: ” for easy filtering.

Editing Prompts

1. Go to Settings → Manage Prompts
2. Find the prompt to edit
3. Click the edit icon
4. Modify content
5. Save changes

Deleting Prompts

1. Go to Settings → Manage Prompts
2. Find the prompt to delete
3. Click the delete icon
4. Confirm deletion

Deleting a prompt cannot be undone. Export important prompts as backup.

Exporting/Importing Prompts

Currently, prompts are stored in browser settings. To backup:

1. Go to Settings → About
2. Export all settings (includes prompts)
3. Save the export file
4. Import on another browser/device to restore

Best Practices

Be Specific

Detailed prompts produce better results. Instead of “You are helpful”, try “You are a helpful Python expert specializing in data science and machine learning.”

Use Examples

Include examples in your prompts to guide AI behavior. Show the format or style you expect.

Set Constraints

Define what the AI should and shouldn’t do. “Do not provide medical advice” or “Always include code examples.”

Test and Iterate

Refine prompts based on results. Start simple, then add detail as needed.

Variables: Use variables liberally in quick prompts. They make templates flexible and reusable.

Time Context: Include {current_date_time} in system prompts for time-aware responses.

Token Limits: Very long system prompts reduce available tokens for conversation. Keep them focused.

Troubleshooting

Prompt not applying

Solutions:

• Ensure prompt is saved
• Refresh the page
• Start a new chat
• Check prompt is selected in dropdown

Variables not working

Solutions:

• Check syntax: use {variable} not (variable) or [variable]
• Ensure no spaces in variable names: {target_lang} not {target lang}
• Verify you’re using Quick Prompt, not System Prompt

Time variables show wrong time

Solutions:

• Check browser time settings
• Verify timezone is correct
• Refresh the page to update

Next Steps

Knowledge Base

Combine prompts with document context

Internet Search

Use prompts with real-time web data

Configuration Settings

Configure RAG and other settings

Sidebar

Use prompts in the sidebar interface