Jentic solves one of AI’s most dangerous blind spots: insecure agent execution and unmanaged credentials — while letting developers securely connect any agent to 1,500+ APIs and 2,000+ workflows
DUBLIN, IRELAND (August 12, 2025) — Jentic, the Dublin-based AI infrastructure company building an agent capability platform for businesses, has launched a secure execution and credentialing system for AI agents, and opened its private beta launch for participants. The update introduces managed authenticated execution and a new developer-facing interface, giving teams centralized control over how agents access APIs and sensitive credentials.
This launch addresses what Jentic describes as a major industry-wide flaw: today, most AI agents today store credentials in prompts or local code, making them difficult to govern and easy to compromise. Jentic now offers a safer, centralized approach. Developers can connect any agent (built using frameworks like LangChain, Autogen, an agent builder platform such as Amazon AWS Bedrock, Google Vertex AI, Microsoft Copilot studio, or anything else) to Jentic’s hosted infrastructure using Model Context Protocol (MCP) or REST. Agents can get work done by securely using over 1,500 APIs and 2,000+ prebuilt workflows, without exposing secrets in the agent code. Developers can try it now at jentic.com.
This release comes amid growing industry concern over AI security, including a recent incident where shared ChatGPT conversations surfaced in public search results. The episode, while unrelated to Jentic, underscores the industry’s urgent need for tools that prevent unintended access to potentially sensitive information.
“The way most AI agents run today creates real risk,” said Sean Blanchfield, co-founder and CEO at Jentic. “You’ve got secrets flying around in prompts and client-side code, with no way to revoke access or enforce consistent controls. That’s a security problem just waiting to happen. Our latest release at Jentic solves that while instantly tapping your agent into thousands of APIs and workflows. It’s like plugging your agent into the grid.”
With this update, developers gain a full control plane for agent infrastructure:
- A web-based credential vault for secure key storage
- Role-based access controls for agents and workflows
- Hosted execution over the open Model Context Protocol (MCP)
- Support for both code-first and no-code agents
- Audit logs to monitor usage and enforce governance
Jentic also flags “Jentic Tested” APIs, highlighting integrations that have been vetted and confirmed to work reliably. Developers are encouraged to contribute improvements via Jentic’s open source repos.
This is the first in a series of infrastructure upgrades that will move Jentic from a developer tool to a full enterprise platform for agents in business. A hosted version of Jentic’s own agent for non-developers is currently in private beta at jentic.com/waitlist.
Access our Press Page and Press Kit for high-res images
About Jentic
Founded in late 2024 and backed by $4.5M in pre-seed funding, Jentic is building AI agents for work. The platform operates at the API layer, not the browser, connecting agents to thousands of services using open standards, secure credential management, and Just-In-Time Tooling. Jentic is open source and based in Dublin, Ireland.
