Descles Released AI Agent Governance Platform
The new control plane allows administrators to manage security and policy constraints for autonomous AI agents.
Updated on Sept. 26, 2026 in Artificial Intelligence
Live Poll
Do you believe companies should require human approval for all AI-driven actions?
Descles has launched a control plane designed to provide governance for AI agents. The platform acts as an intermediary layer that logs activities, enforces access controls, and provides identity attribution for agent requests.
Why it matters
As AI agents increasingly execute complex tasks, organizations require tools to manage behavior and reduce potential security or compliance risks. Descles helps administrators define and enforce policies to ensure these systems function within prescribed parameters.
The Descles platform supports the model-context protocol (MCP) and integrates directly with the OpenAI SDK, Anthropic SDK, and Claude Code. The system currently offers a free tier limited to 1,000 governed requests.
The players
Descles
Descles is an intermediary software platform that provides governance and control mechanisms for autonomous artificial intelligence agents.
OpenAI
OpenAI is an artificial intelligence research laboratory that develops advanced large language models and provides the SDK integrated into the Descles platform.
Anthropic
Anthropic is an AI safety and research company that develops the Claude series of models and provides the SDKs used with Descles.
The details
The platform functions by intercepting tool calls and requests to ensure compliance with defined policies. Administrators can manage these constraints and monitor agent activity, including approvals and denials, through a unified console or API.
Timeline
September 26, 2026: The Descles platform was officially released.
The Tech Race
The rise of Descles mirrors the growing industry shift toward standardizing communication between AI agents and external tools, such as the Model-Context Protocol (MCP). By formalizing how agents interact with infrastructure, Descles positions itself against legacy security tools that lack native support for autonomous agent workflows.
Users can utilize the free tier to govern up to 1,000 AI agent requests, making the tool accessible for testing and small-scale deployment. Developers integrating agents with OpenAI or Anthropic SDKs can use the console to audit agent tool calls and enforce security policies.
The takeaway
Effective AI governance requires moving beyond simple model access to controlling the specific tools and data agents can interact with. Organizations should implement these layers early to maintain security as autonomous agent capabilities expand.
Further reading
For more information on the evolving standards for agentic systems, visit the Artificial Intelligence section.
Source note: This article includes information reported by Dynamic Business.
Live Poll
Do you believe companies should require human approval for all AI-driven actions?
