Description
DKnownAI Guard is a specialized AI agent security API that focuses on detecting prompt injection, jailbreak attempts, and system prompt extraction. It is designed to protect AI agents that execute actions, ensuring they remain secure against manipulation while still functioning effectively. The API classifies requests into four distinct risk categories: AGENT_HACK, SYS_FLAG, CONTENT_FLAG, and SAFE. This classification system allows developers to respond appropriately to various types of requests, enhancing the security of agentic workflows.
Read More
The core functionality of DKnownAI Guard revolves around its ability to detect manipulation attempts targeting AI agents. Unlike traditional content moderation systems that primarily focus on harmful text, DKnownAI Guard prioritizes identifying deceptive tactics aimed at hijacking agent behavior. This proactive approach enables developers to block malicious requests without disrupting legitimate workflows. For instance, if a request attempts to manipulate the agent by issuing commands that override its intended controls, DKnownAI Guard will classify it as AGENT_HACK and recommend immediate blocking.
In addition to its manipulation detection capabilities, DKnownAI Guard also provides context-aware detection. By optionally including conversation context in requests, the API can deliver more accurate classifications, minimizing latency impact. This feature is particularly beneficial for applications where AI agents interact in multi-turn conversations, allowing for a more nuanced understanding of intent and risk.
DKnownAI Guard is built for the Agent Era, where AI agents are expected to execute code, access files, and manage real systems. It supports over 100 languages and integrates seamlessly with various agent frameworks, such as OpenClaw and Claude Code. The API is designed to maintain the autonomy of AI agents while ensuring they are safeguarded against manipulation and operational risks. Developers can start with a free tier that allows for 1,000 API calls per month, making it accessible for testing and initial implementation. As needs grow, a Pro plan is available for 5,000 API calls per month, providing priority support and additional features.
Overall, DKnownAI Guard represents a significant advancement in AI security, focusing on the unique challenges posed by agentic AI systems. By prioritizing manipulation detection and providing clear risk classifications, it empowers developers to create safer and more reliable AI applications.
- CategorySecurity & Privacy
- Pricing ModelFreemium
- DeploymentCloud
- PlatformsWeb
DKnownAI Security API's Core Features
Prompt injection detection
Jailbreak defense
System prompt extraction protection
Four-level risk classification
Multilingual content detection
Agentic framework integration
Context-aware detection
Free tier available
How to use DKnownAI Security API?
Configure: Set up your API key in the Authorization header.
Integrate: Use the API in your agentic workflows for prompt injection detection.
Classify: Send requests to the API for classification of user inputs.
Respond: Act based on the classification results provided by the API.
DKnownAI Security API's Use Cases
- Agent Security
- Risk Classification
- Multi-Turn Conversations
- Compliance Monitoring
- Operational Assistance
