> For the complete documentation index, see [llms.txt](https://ai-security-docs.akto.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ai-security-docs.akto.io/akto-argus-agentic-ai-security-for-homegrown-ai/agentic-ai-discovery/how-to/create-a-custom-data-type.md).

# Create a Custom Data Type

Akto comes with 50+ built-in data types. Create custom types for business-specific patterns in your agent systems.

## Common Agentic Custom Data Types

* Transaction IDs (e.g., "TXN\_12345")
* Agent IDs or session tokens
* Custom API key formats
* Tool invocation signatures
* Context window identifiers

## How to Create

1. Go to **Agentic Discovery > Sensitive Data**
2. Click **Create Custom Data Type**
3. Enter name (e.g., "AGENT\_SESSION\_ID")
4. Define key and value conditions
5. Save

## Add Key Conditions

Define how Akto should detect the key:

* **Equals to**: Exact match (e.g., key = "agent\_id")
* **Matches regex**: Pattern match (e.g., key matches `agent_.*`)
* **Contains**: Partial match

## Add Value Conditions

Define how Akto should detect the value:

* **Matches regex**: Pattern for value (e.g., `\bagent_[a-z0-9]{16}\b`)
* **Equals to**: Exact value match
* **Length**: Value length constraints

## AktoGPT

Use AktoGPT to generate regex patterns automatically:

1. Click **AktoGPT** button
2. Describe your pattern (e.g., "agent session ID with prefix")
3. Review generated regex
4. Add to Akto

## Example: Agent Session ID

**Key Condition**: Equals to "agent\_session" **Value Condition**: Matches regex `\bagt_[a-f0-9]{32}\b`

This detects patterns like: `agent_session: "agt_a1b2c3d4..."`

## Best Practices

* Test regex patterns before adding
* Document custom data types for your team
* Review detected instances to verify accuracy
* Use AktoGPT for complex patterns