Natural Language Queries

OstrichDB’s Natural Language Processing (NLP) feature allows you to interact with your database using plain English commands. This powerful AI-driven interface makes database operations accessible to users of all technical levels.

Overview

What is NLP in OstrichDB?

The NLP feature transforms database operations into conversational interactions:

Plain English Commands: No need to learn complex query syntax
AI Understanding: Advanced language processing interprets your intent
Safe Operations: Review and confirm before execution
Batch Operations: Perform multiple operations in a single request
Learning System: Improves understanding over time

Availability by Plan

Free Tier:

❌ Not available

Pro Tier ($25/month):

✅ 50 requests per day
✅ Basic NLP operations
✅ Standard response time

Enterprise Tier:

✅ Unlimited requests
✅ Advanced NLP operations
✅ Priority processing
✅ Custom training options

Accessing NLP

Getting to the NLP Interface

Navigate to any project in your dashboard
Click “Natural Language Query Processor” in the sidebar
Or use the direct NLP link in project navigation
Select your target project if prompted

Interface Layout

The NLP interface includes:

Query Input Box: Where you type your natural language request
Context Display: Shows current project/collection context
AI Response Area: Displays the AI’s interpretation of your request
Confirmation Panel: Review and approve operations before execution
History Panel: View previous NLP interactions
Usage Counter: Track your daily request limit (Pro tier)

Using Natural Language Queries

Basic Query Structure

NLP queries can be conversational and flexible:

Simple Operations:

“Create a new collection called ‘user-data’”
“Delete the ‘temp’ cluster”
“Show me all records in the users collection”
“Find records where status is active”

Complex Operations:

“Create a user profile collection with name, email, and registration date fields”
“Delete all records in the temporary cluster that are older than 30 days”
“Show me users created this month sorted by registration date”

Supported Operations

Data Creation

"Create a new collection called 'products'"
"Add a cluster named 'electronics' to the products collection"
"Create a record for a laptop with price 999.99"
"Make a new project for my blog application"

Data Retrieval

"Show me all collections in this project"
"List records in the users cluster"
"Find all products with price greater than 100"
"Get the total count of active users"

Data Modification

"Update the price of the laptop to 899.99"
"Change the status of user john_doe to inactive"
"Rename the 'temp' collection to 'archived'"
"Move all draft posts to published status"

Data Deletion

"Delete the old_data collection"
"Remove all records from the temp cluster"
"Delete users who haven't logged in for 6 months"
"Clear the cache cluster"

Batch Operations

"Create collections for users, products, and orders"
"Delete all temporary data from last month"
"Update all draft posts to published and set today's date"
"Create user records for Alice, Bob, and Charlie with default settings"

NLP Workflow

Step-by-Step Process

Enter Your Request: Type your command in natural language
AI Processing: The system interprets your intent
Review Interpretation: Check the AI’s understanding
Modify if Needed: Clarify or correct the interpretation
Confirm Execution: Click “Confirm & Execute” to proceed
View Results: See the outcome of your operation

Example Workflow

User Input:

"Create a new collection for storing customer information with fields for name, email, phone, and registration date"

AI Interpretation:

I understand you want to:
1. Create a new collection named "customer_information"
2. Create a cluster within this collection (default name: "main")
3. Create the following record templates:
   - name (type: STRING)
   - email (type: STRING)
   - phone (type: STRING)
   - registration_date (type: DATE)

Is this correct?

User Confirmation:

✅ Confirm & Execute
❌ Cancel
🔄 Modify Request

Advanced NLP Features

Context Awareness

The NLP system maintains context throughout your session:

Current Project: Remembers which project you’re working in
Recent Operations: References previous operations in the conversation
Data Relationships: Understands connections between your data structures
User Preferences: Learns your naming conventions and patterns

Smart Suggestions

The AI provides intelligent suggestions:

Data Type Recommendations: Suggests appropriate types for your data
Naming Conventions: Follows patterns from your existing data
Relationship Detection: Identifies potential data relationships
Optimization Hints: Suggests performance improvements

Error Handling and Clarification

When the AI is uncertain:

Clarification Questions: Asks for more specific information
Multiple Options: Presents different interpretations for you to choose
Safe Defaults: Uses conservative defaults when unsure
Rollback Options: Offers to undo operations if results aren’t expected

Best Practices

Writing Effective Queries

Be Specific:

❌ “Delete some old data”
✅ “Delete records from the temp cluster older than 30 days”

Use Clear Names:

❌ “Create a thing for users”
✅ “Create a collection called ‘user_profiles’”

Specify Data Types:

❌ “Add a field for age”
✅ “Add an age field as an integer”

Include Context:

❌ “Update the status”
✅ “Update the status field in the users collection to ‘active‘“

Safety Practices

Review Before Execution:

Always check the AI’s interpretation
Verify the scope of operations (especially deletions)
Confirm data types and field names
Check for unintended side effects

Start Small:

Test operations on small datasets first
Use specific rather than broad commands
Create backups before major changes
Practice with non-critical data

Monitor Results:

Check operation outcomes immediately
Verify data integrity after changes
Monitor performance impact of bulk operations
Keep track of your daily usage limits

Troubleshooting NLP

Common Issues

“I don’t understand that request”

Solution: Be more specific and use clearer language
Example: Instead of “fix the data,” try “update all null values in the name field to ‘Unknown’”

“Ambiguous operation detected”

Solution: Provide more context about which data you’re referring to
Example: Specify collection and cluster names explicitly

“Operation would affect too much data”

Solution: Break down large operations into smaller chunks
Example: Instead of “delete old data,” specify “delete records from July 2023 in the logs cluster”

“Daily limit reached”

Solution: Wait for the daily reset or upgrade to Enterprise
Alternative: Use the manual query editor for additional operations

Getting Better Results

Learn from Examples:

Review successful queries in your history
Note which phrasings work best for your use cases
Build a personal library of effective commands

Use Consistent Terminology:

Stick to the same names for similar concepts
Use your established naming conventions
Be consistent with data type references

Provide Feedback:

Confirm when the AI gets it right
Correct misunderstandings clearly
Help improve the system’s understanding of your needs

NLP Limitations

Current Limitations

Complex Queries:

Very complex multi-step operations may need to be broken down
Advanced filtering requires specific syntax
Some edge cases may need manual query writing

Data Analysis:

Statistical operations are limited
Complex aggregations may not be supported
Advanced reporting requires API or manual queries

Bulk Operations:

Large datasets may have processing time limits
Memory constraints for very large operations
Some bulk operations may need to be batched

When to Use Manual Queries

Consider the manual query editor for:

Performance-critical operations
Complex conditional logic
Advanced filtering and sorting
Bulk operations on large datasets
Operations requiring precise control

Usage Monitoring

Tracking Your Usage

Pro Tier Monitoring:

Daily Counter: Shows requests used out of 50 daily limit
Reset Time: Displays when your limit resets (midnight UTC)
Usage History: Track patterns in your NLP usage
Efficiency Metrics: See average operations per request

Enterprise Tier Monitoring:

Unlimited Usage: No daily limits
Performance Metrics: Response times and processing statistics
Usage Analytics: Detailed reports on NLP utilization
Custom Limits: Set organizational limits if desired

Next Steps

To get the most out of NLP:

Practice with Examples: Try the sample queries provided
Explore Your Data: Use NLP to understand your existing data structure
Combine with Manual Queries: Use both interfaces for different needs
Learn from History: Review your successful queries for patterns
Consider Upgrading: Get unlimited access with Enterprise tier

For more advanced database operations:

Manual Query Editor - Direct command interface
Cluster Editor - Visual data management
API Reference - Programmatic access