Catalyst 2.0, the new product from Diagrid, adds durable and verifiable execution to AI agents across major frameworks without requiring any rewrite of application code.

Deployment options span cloud, on-premises, and air-gapped environments, enabling organizations to run agent workloads in diverse settings.

Availability is immediate, with resources to learn more, start using Catalyst, read related blog content, and join a webinar.

Industry endorsements from Diagrid CEO Mark Fussell and ZEISS Group, via Wendelin Niesl, highlight trust, durability, and long-term platform viability.

Diagrid positions itself as the maintainer of Dapr, focusing on secure, auditable, and long-lived agent workflows for regulated sectors like financial services and healthcare.

Supported frameworks include LangGraph, LangGraph Deep Agents, Microsoft Agent Framework, Google ADK, AWS Strands, OpenAI Agents SDK, Claude Managed Agents, CrewAI, Pydantic AI Agents, and Dapr Agents, among others.

Catalyst 2.0 targets durability and trust for production AI, offering cryptographic attestation, execution lineage propagation, and workflow attestation to meet security and compliance needs.

Catalyst 2.0 enables agents to automatically resume from the exact point of failure, with each step cryptographically signed and traceable to its source, supporting multi-cloud and fully air-gapped environments.

The release positions Catalyst 2.0 as a unifying reliability layer that lets teams keep their existing frameworks while gaining durability and verifiability for production use.

Catalyst 2.0 can improve performance of open-source Dapr by up to 10x and supports millions of concurrent agent workflows without re-architecting existing agents.