New open-source platform provides scale, cryptographic identity, permissions, and tamper-proof audit trails for autonomous software, replacing “god-mode” API keys.
TORONTO, ON / ACCESS Newswire / December 10, 2025 / AgentField today announced that it has emerged from stealth to do for AI agents what Kubernetes did for containers: turn them into scalable, governed production infrastructure. The new open-source platform solves the “coordination crisis” of autonomous software by assigning every agent a cryptographic identity, ensuring that when AI makes decisions, like spending money or touching data, it does so with enforceable permissions and mathematical proof of authority.
While designed for any developer deploying autonomous software, this foundation is particularly critical for high-stakes, high-throughput industries-finance, energy, healthcare, logistics, insurance, and critical infrastructure-where small errors can escalate instantly. Across the board, AI is moving out of the chatbot UI and into core systems: reading ledgers, reconciling transactions, and scheduling freight. However, traditional infrastructure cannot support this shift.
“Autonomous software turns your infrastructure into a digital workforce,” said Oktay Goktas, CEO of AgentField. “That workforce doesn’t log in once a day. It runs constantly in the background, fans out across hundreds of APIs, and executes critical operations at machine speed. You cannot manage millions of autonomous decisions with human-centric SSO or brittle API keys. The identity and orchestration layer has to be rebuilt for autonomous software.”
The “Coordination Crisis” of Autonomous Software
When an agent delegates a task, such as moving funds, updating a patient record, or re-routing a shipment, it triggers a chain of downstream actions where a single request cascades into dozens of autonomous decisions. Traditional authorization systems (OAuth, IAM) fail here because they were built for deterministic human workflows; relying on synchronous checks or broad API keys for software that reasons and acts autonomously causes the entire trust model to collapse.
AgentField solves this by combining Kubernetes-style orchestration with decentralized identity:
-
Cryptographic Identity (DIDs): Every agent receives a W3C Decentralized Identifier.
-
Verifiable Credentials: Agents issue tamper-proof proofs when delegating tasks. A supply chain agent can cryptographically prove it was authorized by a specific risk policy, and that its authority traces back through a chain of agents to an original human or system decision, without needing a central callback.
-
Tamper-Proof Audits: Security teams get cryptographic receipts for every action, not just text logs.
“We are watching old assumptions break in real time,” said Santosh Kumar Radha, CTO of AgentField. “When a $250,000 transfer happens at 3:17 a.m., you need to know exactly who authorized it and under what policy. AgentField ensures that even if an agent functions autonomously five hops down a chain, its authority can be mathematically verified.”
Production Infrastructure for Thinking Software
AgentField provides the control plane for autonomous agents the way Kubernetes provides one for containers, but for identity-bearing, long-running intelligence. Beyond identity, AgentField provides the operational primitives needed to run stateful agents in production:
-
Long-running execution: Support for agents that work on tasks for minutes or hours.
-
Async orchestration: Built-in retries, webhook-based triggers, and backpressure handling for multi-agent workflows.
-
Policy as Code: Access control and IAM policies are enforced by the control plane instead of prompts or ad-hoc scripts.
Most teams today silently reinvent this backend plumbing with brittle scripts and ad-hoc glue code. AgentField provides it as a coherent, production-grade control plane where trust travels with the agent across systems and organizations, an essential property of a real agent economy.
Open-Source Strategy & Availability
AgentField is launching as an Apache 2.0 open-source project available today on GitHub. The founders view open standards as critical for an “agent economy” where software from different vendors must coordinate trustlessly.
Backed by Panache Ventures and Brightspark Ventures, AgentField was founded by repeat entrepreneurs and PhD holders who previously built Agnostiq and its open-source platform Covalent, acquired by DataRobot in February 2025.
For more information, visit agentfield.ai.
ABOUT AGENTFIELD:
AgentField provides identity and scale infrastructure for autonomous software. Based in Toronto, the company builds the open-source foundations that give AI agents cryptographic identities, delegatable permissions, and production-grade orchestration. Founded by Oktay Goktas, PhD (CEO), and Santosh Kumar Radha, PhD (CTO), who previously built and sold Agnostiq to DataRobot in 2025. Learn more at agentfield.ai.
MEDIA CONTACT: Nina Pfister, MAG PR at nina@mooringadvisorygroup.com
SOURCE: AgentField
View the original press release on ACCESS Newswire
