Diagrid brings cryptographic proof to AI agent and workflow execution

Overview

Diagrid released Dapr 1.18, an update to the open-source runtime that lets organizations cryptographically prove how an AI agent or workflow executed, who held custody of the work, and whether its history was altered. The release centers on a capability called verifiable execution, built from three new features. Diagrid is the primary contributor to Dapr, a graduated project of the Cloud Native Computing Foundation. The update also includes other changes such as a stable Jobs API and general availability of hot reloading.

Key Takeaways

Diagrid released Dapr 1.18, adding a capability called verifiable execution.
Verifiable execution lets organizations cryptographically prove how an AI agent or workflow ran and whether its history was altered.
It is built from three features: Workflow History Signing, Workflow History Propagation, and Workflow Attestation.
Workflow History Signing is backed by application identities tied to the open SPIFFE standard.
Diagrid is the primary contributor to Dapr, a graduated project of the Cloud Native Computing Foundation.

Stats & Key Facts

#Diagrid raised $20 million in a Series A round
#Hot reloading reaches general availability across six resource types
#Diagrid was founded in 2021

Diagrid brings cryptographic proof to AI agent and workflow execution

What Verifiable Execution Provides

The release centers on proving what autonomous systems did.

›It lets organizations cryptographically prove how an agent or workflow executed.
›It shows who held custody of the work and whether its history was altered.
›It gives security and compliance teams a chain of custody for decisions autonomous systems make.

The pitch comes as enterprises move AI agents into production and face accountability questions that resilience alone does not answer. Distributed systems have long recovered from failures, but verifying exactly what happened during execution has been harder.

The Three New Features

Verifiable execution is built from three features.

›Workflow History Signing signs execution records so they become tamper-evident and independently verifiable, backed by application identities tied to the open SPIFFE standard.
›Workflow History Propagation lets execution lineage travel across workflow, service, and application boundaries.
›Workflow Attestation passes verified execution context to activities and child workflows so policy and compliance decisions rely on confirmed provenance.

Why It Matters Now

The release targets accountability for AI agents in production.

›Yaron Schneider, Diagrid co-founder and CTO, said the first wave of AI focused on making models intelligent and the next will focus on making AI systems trustworthy.
›He said organizations need to prove what happened when an agent approves a transaction, accesses sensitive data, or triggers a business process.
›That proof needs to be tamper-proof.

Schneider chairs the Agentic AI Foundation Workflows Working Group, and the features were designed and contributed by Diagrid engineers with the Dapr community.

Other Changes in Dapr 1.18

The update includes several additional changes.

›The Jobs API for scheduling future and recurring work graduates to stable.
›Hot reloading for components and configurations reaches general availability across six resource types, enabling updates without restarts.
›Actor applications can open a single bidirectional gRPC stream to the sidecar to receive callbacks without exposing an inbound server port, reducing attack surface.

The runtime also gains IPv6 and dual-stack support.

About Diagrid and Dapr

Diagrid leads the Dapr project and sells a managed platform.

›The creators of Dapr founded Diagrid in 2021, and the company raised $20 million in a Series A round at its 2022 launch.
›Diagrid sells Catalyst, a managed platform for long-lived agent workloads with built-in identity, access controls, and governance.
›Its customers include companies in financial services and healthcare.

Dapr 1.18 is available now as an open-source release and on Diagrid Catalyst Cloud, with additional Catalyst support due later this month.

Frequently Asked Questions

What is verifiable execution in Dapr 1.18?

It is a capability that lets organizations cryptographically prove how an AI agent or workflow executed, who held custody of the work, and whether its history was altered.

What three features make up verifiable execution?

Workflow History Signing, Workflow History Propagation, and Workflow Attestation.

What standard backs the signing feature?

Workflow History Signing is backed by application identities tied to the open SPIFFE standard.

What other changes does Dapr 1.18 include?

The Jobs API graduates to stable, hot reloading reaches general availability across six resource types, actor apps can use a single bidirectional gRPC stream to reduce attack surface, and the runtime gains IPv6 and dual-stack support.

What is Diagrid's relationship to Dapr?

Diagrid is the primary contributor to Dapr, a graduated project of the Cloud Native Computing Foundation, and its creators founded Diagrid in 2021.

Dapr 1.18 adds verifiable execution so teams can cryptographically prove what their AI agents and workflows actually did.