Custom Prompts

Create reusable AI prompts using database pages as sources

Overview

Custom Prompts let you create your own AI prompt library by loading database pages as sources for AI interactions. This feature enables you to build standardized, reusable prompts that incorporate your own content, ensuring consistent AI responses across your team.

Custom Prompts were introduced in v0.9.3 and work with both AI Chat and the Ask AI feature in documents.

What Are Custom Prompts?

Custom Prompts turn your database pages into AI prompt templates:

• Database-powered - Store prompts in structured database format
• Reusable - Use the same prompt multiple times across different contexts
• Customizable - Tailor prompts to your specific workflows and needs
• Team-wide - Share prompt libraries with your team members

Creating Custom Prompts

1. Create a Database

   Create a new database (grid, board, or calendar view) to store your prompts.

2. Add Prompt Fields

   Set up your database with these essential properties:

   • Prompt Name (Text) - The title/name of your prompt
   • Prompt Content (Text) - The actual prompt text
   • Category (Select) - Optional categorization (Writing, Coding, Analysis, etc.)
   • Tags (Multi-select) - Optional tags for filtering
3. Add Your Prompts

   Create rows in the database, each representing one prompt:

   Name: "Meeting Summary"
   Content: "Summarize this meeting note into key points, decisions, and action items with owners."
   Category: "Productivity"
   
4. Load as Custom Prompts

   In AI Chat or Ask AI:

   1. Click the prompt selector icon
   2. Choose Load Custom Prompts
   3. Select your database page
   4. Your prompts are now available for use

Using Custom Prompts

In AI Chat

1. Open Prompt Menu

   Click the prompt icon in the chat input area or type / to open the prompt selector.

2. Select Your Custom Prompt

   Browse the prompt library and select your custom prompt from the list.

3. Apply the Prompt

   The prompt content is applied to your message. Add any additional context or variables.

4. Send to AI

   Press Enter to send the customized prompt to the AI.

In Document Ask AI

1. Select text in your document
2. Open the Ask AI menu (Cmd/Ctrl + /)
3. Choose a custom prompt from your library
4. The AI applies the prompt to your selected content

You can load multiple database pages as custom prompt sources, organizing prompts by project, team, or use case.

Custom Prompt Examples

Writing Prompts

Content Improvement

Name: Improve Clarity

Prompt:

Review this text and improve its clarity and readability. 
Make it more concise while preserving all key information. 
Use simple language and short sentences.

Tone Adjustment

Name: Make Professional

Prompt:

Rewrite this text in a professional tone suitable for 
business communication. Maintain the core message but 
adjust language, formality, and structure.

Format Conversion

Name: Convert to Email

Prompt:

Convert these notes into a professional email format with 
a clear subject line, greeting, structured body, and 
appropriate closing.

Analysis Prompts

Meeting Summary

Name: Summarize Meeting

Prompt:

Analyze these meeting notes and provide:
1. Key discussion points
2. Decisions made with rationale
3. Action items with assigned owners
4. Unresolved questions or concerns

Document Review

Name: Review Proposal

Prompt:

Review this proposal and provide:
- Strengths and compelling points
- Potential weaknesses or gaps
- Questions that should be addressed
- Suggestions for improvement

Development Prompts

Code Review

Name: Code Review Checklist

Prompt:

Review this code for:
- Code quality and readability
- Potential bugs or errors
- Performance considerations
- Best practices and patterns
Provide specific suggestions for improvement.

API Documentation

Name: Generate API Docs

Prompt:

Generate API documentation for this endpoint including:
- Description and purpose
- Request parameters and types
- Response format and examples
- Error codes and handling
- Usage examples in cURL and JavaScript

Advanced Custom Prompts

Parameterized Prompts

Create prompts with placeholders for variables:

Name: "Product Feature Description"
Prompt: "Write a product feature description for [FEATURE_NAME] 
that highlights [KEY_BENEFIT] and targets [AUDIENCE]. 
Include a compelling headline and 2-3 supporting bullet points."

When using this prompt, replace the bracketed placeholders with specific values.

Multi-Step Prompts

Create complex prompts with multiple stages:

Name: "Comprehensive Content Review"
Prompt: 
"Step 1: Read the content and identify the main purpose and audience.
Step 2: Evaluate clarity, structure, and readability.
Step 3: Check for factual accuracy and completeness.
Step 4: Provide a summary with specific improvement recommendations."

Context-Rich Prompts

Include context and constraints in your prompts:

Name: "User Story Generation"
Prompt: "Create user stories following this format:
'As a [user type], I want to [action], so that [benefit].'

Context:
- Follow Agile/Scrum best practices
- Include acceptance criteria
- Estimate story points (1, 2, 3, 5, 8)
- Identify dependencies or risks"

Organizing Your Prompt Library

Database Structure

Optimal database setup for Custom Prompts:

Property	Type	Purpose
Name	Text	Prompt identifier
Prompt	Text	The actual prompt content
Category	Select	Group by type (Writing, Code, Analysis)
Use Case	Multi-select	Specific applications
Model	Select	Best AI model for this prompt
Variables	Text	List of placeholders to fill
Version	Number	Track prompt iterations
Last Used	Date	Track usage frequency

Categories to Consider

Content Creation

Writing, editing, formatting, and content improvement prompts

Analysis

Summarization, review, comparison, and evaluation prompts

Development

Code review, documentation, debugging, and testing prompts

Planning

Project planning, brainstorming, and strategy prompts

Using Views for Organization

Create different database views for prompt organization:

• By Category - Board view grouped by Category
• By Frequency - Sort by Last Used date
• By Project - Filter prompts for specific projects
• Favorites - Filter for most-used prompts

Sharing Custom Prompts

Team Prompt Libraries

Share your prompt database with team members:

1. Set appropriate sharing permissions on the database page
2. Team members can load the shared database as custom prompts
3. Everyone uses the same standardized prompts
4. Update prompts centrally for entire team

When you update prompts in the database, changes are immediately available to everyone using that database as a custom prompt source.

Project-Specific Prompts

Create separate prompt databases for different projects:

• Marketing Prompts - Social media, blog posts, campaigns
• Engineering Prompts - Code review, documentation, architecture
• Product Prompts - User stories, feature specs, requirements
• Support Prompts - Response templates, troubleshooting guides

Best Practices

Be Specific

Clear, detailed prompts produce better AI responses. Include format, tone, and length requirements.

Test and Iterate

Refine prompts based on AI outputs. Track versions to compare effectiveness.

Include Examples

Add example outputs in your prompts to guide the AI’s response format.

Document Variables

List all placeholders and variables that need to be filled when using the prompt.

Writing Effective Prompts

1. Define the Task

   Start with a clear action verb: “Summarize”, “Analyze”, “Generate”, “Review”

2. Provide Context

   Include relevant background, constraints, and requirements

3. Specify Format

   Define how the output should be structured (bullets, paragraphs, tables)

4. Set Tone and Style

   Indicate the desired tone: professional, casual, technical, etc.

5. Add Examples

   Show what good output looks like with concrete examples

Troubleshooting

Prompts Not Appearing

If your custom prompts don’t show up:

1. Verify the database page is loaded as a custom prompt source
2. Check that the database has the correct property names
3. Ensure the database has at least one row with prompt content
4. Reload the custom prompt source

Inconsistent AI Responses

If the AI gives varying results with the same prompt:

• Make the prompt more specific and detailed
• Include examples of desired outputs
• Specify format and structure requirements
• Mention what to avoid or exclude

Managing Multiple Prompt Libraries

• Use clear naming for database pages: “Team Prompts - Marketing”
• Unload unused prompt libraries to reduce clutter
• Create a master database linking to all prompt libraries
• Document which prompts work best with which AI models

Custom Prompts vs. AppFlowy Prompt Library

Feature	Custom Prompts	Built-in Library
Source	Your database pages	Pre-built by AppFlowy
Customization	Fully customizable	Fixed prompts
Sharing	Share with team	Available to all users
Updates	You control updates	Updated by AppFlowy
Use Case	Organization-specific	General purpose

Use both! Leverage the built-in library for common tasks and create custom prompts for your specific workflows.

Related Features

AI Chat

Use custom prompts in interactive AI conversations

Databases

Learn more about creating and organizing databases

Templates

Create reusable document and database templates

Collaboration

Share prompt libraries with your team