> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pylar.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Why Pylar?

> Discover the benefits and advantages of using Pylar for secure AI agent data access

## Why Choose Pylar?

Pylar solves critical challenges in AI agent development by providing a secure, governed data access layer. Here's why organizations choose Pylar over building custom solutions or giving agents direct database access.

## Security & Governance First

### Fine-Grained Access Control

Unlike direct database connections, Pylar lets you define exactly what data agents can access at the view, column, and row level. This means you can:

* **Restrict Sensitive Data**: Automatically mask or exclude PII, financial data, or other sensitive information
* **Row-Level Security**: Control access based on user context, regions, or business rules
* **Query Validation**: Prevent unauthorized queries before they reach your database
* **Audit Every Access**: Track who accessed what data and when

<Tip>
  With Pylar, you can start with read-only access to non-sensitive data and gradually expand as you build confidence in your governance policies.
</Tip>

### Compliance Ready

Pylar helps you meet regulatory requirements out of the box:

* **Complete Audit Trails**: Every query is logged with full context
* **Data Lineage**: Understand where data came from and how it's used
* **Access Controls**: Enforce policies that meet SOC 2, GDPR, HIPAA, and other standards
* **Documentation**: Auto-generated documentation for compliance reviews

<Warning>
  Direct database access makes compliance nearly impossible. Pylar provides the governance layer you need.
</Warning>

## Developer Experience

### AI-Powered Tool Creation

Instead of writing custom integrations, use natural language to create MCP tools:

* **Natural Language Prompts**: Write "create a tool to fetch error codes, events, and deal information" and Pylar's AI configures it
* **Manual Control Available**: Fine-tune tools manually when needed
* **Multiple Tools Per View**: Create different tools on the same view for flexibility
* **Test Before Deploy**: Verify tools work correctly before publishing
* **Standard Protocol**: Uses MCP, the emerging standard for AI tooling

### One View, Multiple Tools

Create multiple MCP tools on a single view, each governing how agents interact with your data differently. All tools work with any MCP-compatible agent:

* Claude Desktop
* Cursor
* Windsurf
* Custom agent frameworks
* Any MCP-compatible tool

<Check>
  Write your data access logic once, create multiple tools, use everywhere.
</Check>

### Rapid Iteration Without Redeployment

* **Adjust Anytime**: Update queries, view scope, or tools on the platform
* **Automatic Propagation**: Changes reflect everywhere automatically—no redeployment needed
* **Test Before Publish**: Test MCP tools before deploying to ensure they work
* **Iterative Improvement**: Use Evals insights to refine views and tools continuously

## Operational Excellence

### In-Depth Evals & Observability

Click "Evals" to get comprehensive insights into how agents interact with your views:

* **Interaction Patterns**: See exactly how agents use your tools
* **Success vs. Errors**: Track successful queries and identify failures
* **Error Analysis**: Detailed breakdown—what errors occurred, when, and why
* **Query Shape**: Understand the types of queries agents are making
* **Raw Logs**: Access full query logs for debugging
* **Iterative Optimization**: Use insights to adjust view scope and tool configurations

### Scalability Without Headaches

Pylar scales with you:

* **Multi-Database Support**: Connect to BigQuery, Snowflake, PostgreSQL, and more
* **Cross-Database Joins**: Join data across multiple databases in a single query
* **Project Organization**: Organize views and tools in projects for better management
* **SQL IDE**: Built-in SQL editor for easy query development
* **Views Are the Only Access**: Agents never get raw database access—complete security

<Info>
  Start small with one database and one view. Scale to hundreds of views across multiple data sources as you grow. All within projects that keep everything organized.
</Info>

## Cost Efficiency

### Reduce Development Time

Building custom agent integrations typically takes weeks or months. With Pylar:

* **Days Instead of Months**: Get your first tools running in hours
* **No Maintenance Burden**: Pylar handles protocol updates and compatibility
* **Focus on Value**: Spend time on your business logic, not infrastructure

### Lower Infrastructure Costs

* **Connection Efficiency**: Shared connections reduce database load
* **Query Optimization**: Built-in query analysis and optimization
* **Caching Layer**: Reduce redundant queries and costs
* **Resource Management**: Prevent runaway queries that spike costs

## Real-World Benefits

### For Data Teams

* **Self-Service with SQL IDE**: Use the built-in SQL IDE to create views without waiting for engineering
* **Cross-Database Analysis**: Join data across warehouses without complex pipelines
* **Governance Without Friction**: Maintain control—views are the only access level
* **Evals Insights**: Understand exactly how agents interact with your data
* **Iterative Improvement**: Use evals to continuously refine views and tools

### For Engineering Teams

* **No Custom Integration Code**: AI-powered tool creation eliminates boilerplate
* **Standardization**: One platform for all agent data access needs
* **Faster Shipping**: Get agents to market faster with instant tool creation
* **Easier Debugging**: Evals provide raw logs, query shapes, and error analysis
* **No Redeployment**: Changes reflect automatically everywhere

### For Organizations

* **Complete Isolation**: Views are the only access level—agents never get raw database access
* **Risk Reduction**: Eliminate direct database access for AI agents
* **Faster AI Adoption**: Remove blockers with AI-powered tool creation
* **Compliance Confidence**: Meet regulatory requirements with built-in controls
* **Cost Control**: Monitor and manage data access costs through Evals
* **Iterative Optimization**: Continuously improve based on real agent behavior

## Comparison: Pylar vs. Alternatives

### Pylar vs. Custom Integration

| Feature            | Pylar     | Custom Integration |
| ------------------ | --------- | ------------------ |
| Time to First Tool | Hours     | Weeks/Months       |
| Maintenance        | Automatic | Ongoing            |
| Governance         | Built-in  | Manual             |
| Observability      | Included  | Custom Build       |
| Multi-Database     | Supported | Per-DB Code        |
| Protocol Updates   | Automatic | Manual             |

### Pylar vs. Direct Database Access

| Feature        | Pylar                | Direct Access      |
| -------------- | -------------------- | ------------------ |
| Security       | Fine-grained control | All or nothing     |
| Compliance     | Audit-ready          | Difficult to audit |
| Cost Control   | Built-in monitoring  | No visibility      |
| Error Handling | Managed              | Manual             |
| Scaling        | Automatic            | Manual             |

## When Pylar Makes Sense

Pylar is ideal if you:

* ✅ Have AI agents that need structured data access
* ✅ Need to maintain data governance and compliance
* ✅ Want to avoid writing custom integration code
* ✅ Need visibility into agent data usage
* ✅ Work with multiple databases or data sources
* ✅ Want to scale AI agent usage safely
* ✅ Need to support multiple agent frameworks

## Getting Started

Ready to experience these benefits? Check out our [Quick Start guide](/introduction/quick-start) to get up and running in minutes.

<Card title="Start Building" icon="rocket" href="/introduction/quick-start">
  Get your first Pylar tools running in minutes
</Card>