llms.txt
Compare with other sites' llms.txt files →# Diagrid
> Diagrid Catalyst is a durable execution platform for AI agents and workflows.
Diagrid makes AI agents, workflows and apps durable, verifiable and secure in production, built on open-source Dapr (a CNCF graduated project). Diagrid's platform makes any agent and MCP server crash-safe with guaranteed recovery and safe replay under failure. It provides durable execution, service discovery, identity management, and end-to-end tracing for production AI agent workloads.
## Products
### Diagrid
Diagrid Catalyst is a durable execution platform for AI agents and workflows. It gives agents, workflows, MCP servers and apps durable execution, verifiable execution and a cryptographic identity with policy-based access, and works with any agent framework. It also provides service invocation, pub/sub messaging and state management, without managing infrastructure.
**Key capabilities:**
- **Durable Execution & Workflows**: Build crash-resilient agents and orchestration pipelines with automatic state recovery, guaranteed completion, and safe replay under failure. A managed alternative to self-hosting Temporal or other durable execution engines.
- **Agent & MCP Service Discovery**: Identify, locate, and securely communicate between agents, MCP servers, and microservices.
- **Authentication & Authorization**: Built-in identity and access management with mTLS for all service-to-service communication.
- **Tracing & Observability**: End-to-end distributed tracing across agents and workflows.
- **Workflow Visualizers**: Visual debugging and monitoring of durable workflow execution.
- **Multi-Region Deployment**: Run agents across regions with serverless scale-out.
**Supported agent frameworks:** LangGraph, CrewAI, PydanticAI, Google ADK, AWS Strands, OpenAI Agents, LangChain Deep Agents, Microsoft Agent Framework, Spring AI.
**Framework solutions — take any agent framework to production:**
- [LangGraph Production](https://www.diagrid.io/solutions/langgraph-production)
- [CrewAI Production](https://www.diagrid.io/solutions/crewai-production)
- [PydanticAI Production](https://www.diagrid.io/solutions/pydanticai-production)
- [Google ADK Production](https://www.diagrid.io/solutions/google-adk-production)
- [AWS Strands Production](https://www.diagrid.io/solutions/aws-strands-production)
- [OpenAI Agents Production](https://www.diagrid.io/solutions/openai-agents-production)
- [LangChain Deep Agents Production](https://www.diagrid.io/solutions/langchain-deep-agents-production)
- [Microsoft Agent Framework Production](https://www.diagrid.io/solutions/microsoft-agent-framework-production)
- [Spring AI Production](https://www.diagrid.io/solutions/spring-ai-production)
- [All Solutions](https://www.diagrid.io/solutions)
**Use cases:**
- Agentic AI applications with durable execution
- Process orchestration and workflow automation
- Human-in-the-loop workflows
- Event-driven microservices
- Architecture modernization from legacy to cloud-native
**Pricing:** Free tier (Diagrid Cloud) for non-critical workloads. Enterprise tier with 99.99% SLA, dedicated hosting, and support.
- [Diagrid Overview](https://www.diagrid.io/catalyst)
- [Diagrid Documentation](https://docs.diagrid.io/)
- [Diagrid Quickstarts](https://docs.diagrid.io/getting-started/quickstarts/)
- [Pricing](https://www.diagrid.io/pricing)
### Dapr Ops Dashboard
Dapr Ops Dashboard (formerly Diagrid Conductor) is a SaaS tool that automates the operational management of Dapr on Kubernetes. It handles installation, upgrades, mTLS certificate rotation and visualization of open-source Dapr across clusters, enforces security best practices, and provides monitoring across all Dapr Kubernetes clusters.
**Key capabilities:**
- Automated Dapr installation, upgrades, and patching with rollback
- Production best-practices enforcement and misconfiguration prevention
- Real-time analysis and alerting on 150+ Dapr metrics
- Application dependency visualization
- Automated certificate management with zero downtime
- Multi-cluster management across any Kubernetes distribution
- [Dapr Ops Dashboard Overview](https://www.diagrid.io/dapr-ops-dashboard)
- [Dapr Ops Dashboard Documentation](https://docs.diagrid.io/dapr-open-source/dapr-ops-dashboard/)
### Diagrid Dapr Dev Dashboard
The Dapr Dev Dashboard is a free, open source, single-binary companion for local Dapr development. It discovers the Dapr apps already running on your machine, shows their components and workflows, and streams their logs in one browser window. It is a local development tool only — no authentication, no multi-user support, and not for Kubernetes, staging, or production.
**Key capabilities:**
- Dapr Workflow debugging with the full event history, live status, input/output, and terminate/purge in bulk
- Workflow state read directly from local Redis, PostgreSQL, SQLite, and MongoDB, plus containerized stores via the sidecar gRPC API
- Live table of every discovered Dapr app with health, ports, PIDs, runtime metadata, loaded components, actor types, and pub/sub subscriptions
- Component Builder and Resiliency Builder that generate valid Dapr YAML from a guided form
- Live log tailing for both app and sidecar logs with level coloring and keyword highlighting
- Automatic discovery of dapr run, dapr run -f, .NET Aspire, Docker Compose, and Dapr Testcontainers
- [Dapr Dev Dashboard Overview](https://www.diagrid.io/dev-dashboard)
- [Dapr Dev Dashboard Documentation](https://docs.diagrid.io/develop/local-development/dev-dashboard)
- [Dapr Dev Dashboard Source](https://github.com/diagridio/dev-dashboard)
### Diagrid Dapr Distribution for Enterprise (D3E)
D3E is a custom distribution of open-source Dapr for organizations running Dapr at scale or under strict security review. It provides namespace isolated Dapr control planes for every team in one Kubernetes cluster, on premises or air-gapped, while maintaining 100% compatibility with the open-source Dapr APIs and SDKs. Requires a Premium Dapr support plan.
**Key capabilities:**
- Multi-tenancy: scope a control plane to one namespace or a defined list, and run multiple control planes side by side in one cluster
- Reduced ClusterRole requirements versus open-source Dapr, with granular namespace and infrastructure access control
- A CRD-free and ClusterRole-free installation option for environments that prohibit them
- CVE fixes and critical patches backported to custom Dapr versions not available in open source
- Drop-in replacement installed by Helm chart from the Diagrid registry, with application code and components unchanged
- Four installation options: single namespace isolation, multi-namespace isolation, multiple Dapr installations, and ClusterRole/CRD free
- 24/7 enterprise support with guaranteed response times on production issues
- [D3E Overview](https://www.diagrid.io/diagrid-dapr-distribution)
- [D3E Documentation](https://docs.diagrid.io/deploy/self-hosted-dapr/diagrid-enterprise/d3e/)
### Diagrid Enterprise for Dapr
Enterprise-grade Dapr support for production deployments, including security-enhanced Dapr binaries (D3E), 24/7 production support with 1-hour SLA, CVE resolution, architectural reviews, and expert guidance.
- [Enterprise for Dapr](https://www.diagrid.io/dapr-enterprise)
## Built on Open Source
Diagrid is built on top of these open-source projects:
- [Dapr](https://dapr.io/) — Distributed Application Runtime (CNCF Graduated)
- [SPIFFE](https://spiffe.io/) — Secure Production Identity Framework
- [CloudEvents](https://cloudevents.io/) — Event data specification
- [KEDA](https://keda.sh/) — Kubernetes Event-driven Autoscaling
- [OpenTelemetry](https://opentelemetry.io/) — Observability framework
## Comparisons
- [Diagrid vs Temporal](https://www.diagrid.io/compare-temporal)
- [Diagrid vs Dapr](https://www.diagrid.io/compare-dapr)
- [Dapr vs Orkes Conductor](https://www.diagrid.io/compare-orkes)
- [Dapr vs Camunda](https://www.diagrid.io/compare-camunda)
## Developers
- [Documentation](https://docs.diagrid.io/)
- [What is AI Orchestration?](https://www.diagrid.io/ai-orchestration)
- [Workflow Composer](https://workflows.diagrid.io/)
- [Diagrid Labs (GitHub)](https://github.com/diagrid-labs)
## APIs and agent access
Agents can work with Diagrid over its REST API, through its hosted MCP server, or with the Diagrid CLI. The free tier is self-serve; sign up and get started from the [Diagrid documentation](https://docs.diagrid.io).
- [Diagrid control-plane OpenAPI spec](https://www.diagrid.io/openapi.json): OpenAPI 3 description of the Diagrid control-plane REST API (projects, app identities, components, agents, MCP servers). Requests and responses follow JSON:API.
- [Diagrid control-plane API reference](https://docs.diagrid.io/api/control-plane/catalyst/)
- [Diagrid data-plane (Dapr) OpenAPI spec](https://docs.diagrid.io/api/catalyst-openapi.yaml)
- [API catalog (RFC 9727)](https://www.diagrid.io/.well-known/api-catalog): machine-readable index of the APIs above.
- Diagrid MCP server: https://mcp.cloud.r1.diagrid.io/mcp (Streamable HTTP, OAuth 2.0, scopes catalyst:read and catalyst:write)
- [Diagrid MCP server card](https://www.diagrid.io/.well-known/mcp/server-card.json)
- [Connect an MCP client to Diagrid](https://docs.diagrid.io/develop/mcp/connect/)
- [Install the Diagrid CLI](https://docs.diagrid.io/getting-started/install-cli/)
## Resources
- [Blog](https://www.diagrid.io/blogs)
- [Customer Stories](https://www.diagrid.io/customers)
- [Durable Execution & Dapr University (Learning Paths)](https://www.diagrid.io/university)
- [Make AI Agents Durable (Courses)](https://www.diagrid.io/university/ai)
- [Make Workflows Durable (Courses)](https://www.diagrid.io/university/workflow)
- [Make Applications Durable (Courses)](https://www.diagrid.io/university/dapr)
- [Learn (Guides)](https://www.diagrid.io/learn)
- [FAQ](https://www.diagrid.io/faq)
- [Contact](https://www.diagrid.io/contact)
## Learn
Free, in-depth guides taking you from AI fundamentals to production-ready, secure AI agents.
### AI & Agent Fundamentals
- [AI 101: Generative AI, Agents, and LLMs](https://www.diagrid.io/learn/ai-101-generative-ai-agents-llms)
- [Deep Dive into Agents: Beyond Chatbots and Copilots](https://www.diagrid.io/learn/deep-dive-agents)
- [From Prototype to Production: What Changes for AI Apps](https://www.diagrid.io/learn/prototype-to-production)
- [Agents vs Workflows vs Chains](https://www.diagrid.io/learn/agents-vs-workflows-vs-chains)
- [Model Context Protocol (MCP) Explained](https://www.diagrid.io/learn/model-context-protocol-mcp)
- [Single Agent vs Multi-Agent Systems](https://www.diagrid.io/learn/single-agent-vs-multi-agent-systems)
### Building AI Agents
- [Architecting a Production-Ready AI Agent](https://www.diagrid.io/learn/production-ready-agent)
- [Durable Agents: Surviving Crashes, Restarts, and Long-Running Tasks](https://www.diagrid.io/learn/durable-agents)
- [Understanding AI Agent Patterns: From Workflows to Autonomous Orchestration](https://www.diagrid.io/learn/understanding-ai-agent-patterns)
- [Cost Optimization for LLM-Powered Agents](https://www.diagrid.io/learn/cost-optimization-for-llm-powered-agents)
### Security & Enterprise Readiness
- [Securing AI Agents: Implementing Agent Identity](https://www.diagrid.io/learn/securing-ai-agents-implementing-agent-identity)
- [Data Governance for AI Agents](https://www.diagrid.io/learn/data-governance-for-ai-agents)
- [Handling PII in Agentic Workflows and Applications](https://www.diagrid.io/learn/handling-pii-agentic-workflows)
- [Applying Cryptographic Attestation to Agentic Applications](https://www.diagrid.io/learn/cryptographic-attestation-ai-agents)
### Agent Framework Integrations
- [Integrating Diagrid Catalyst with LangGraph](https://www.diagrid.io/learn/integrating-diagrid-catalyst-langgraph)
- [Using AWS Strands with Diagrid Catalyst Durable Execution](https://www.diagrid.io/learn/aws-strands-durable-execution)
- [Using CrewAI with Catalyst's Durable Execution](https://www.diagrid.io/learn/crewai-durable-execution)
- [Using Google ADK with Catalyst's Durable Execution](https://www.diagrid.io/learn/google-adk-durable-execution)
- [Using Microsoft Agent Framework with Catalyst's Durable Execution](https://www.diagrid.io/learn/microsoft-agent-framework-durable-execution)
- [Using Anthropic's Claude Agent SDK with Catalyst's Durable Execution](https://www.diagrid.io/learn/claude-agent-sdk-durable-execution)
### Observability, Debugging & Reliability
- [Common Failure Modes in Production Agents](https://www.diagrid.io/learn/common-failure-modes-in-production-agents)
- [Tracking Token Usage, Latency, and Cost Across Agents](https://www.diagrid.io/learn/agent-observability-token-usage-latency-cost)
## The Evolution of AI Agents (editorial series)
- [The Evolution of AI Agents](https://www.diagrid.io/special-projects/agentic-execution-evolution): Pillar. Seven stages of the software loop - procedural program, Windows message loop, game loop, Linux event loop, Kubernetes control loop, agent loop, durable agent loop - with a glossary and the definition of agentic durable execution.
- [Before the Loop: When Programs Were Mostly Linear](https://www.diagrid.io/blog/agentic-execution-evolution-1-before-the-loop): Why the linear execution model stopped being enough once the outside world became part of the execution model.
- [The Windows Message Loop](https://www.diagrid.io/blog/agentic-execution-evolution-2-windows-message-loop): GetMessage/DispatchMessage, event-driven programming, and the inversion of control that every agent SDK inherited.
- [The Game Loop](https://www.diagrid.io/blog/agentic-execution-evolution-3-game-loop): Continuous loops, world state carried forward, and the ancestry of stateful AI agents and AI agent state management.
- [The Linux Event Loop](https://www.diagrid.io/blog/agentic-execution-evolution-4-linux-event-loop): select vs poll vs epoll, efficient waiting, and why long running workflows need more than readiness notification.
- [The Kubernetes Control Loop](https://www.diagrid.io/blog/agentic-execution-evolution-5-kubernetes-control-loop): Reconciliation loops, desired state vs actual state, the controller pattern, and how it differs from workflow orchestration.
- [The Agent Loop](https://www.diagrid.io/blog/agentic-execution-evolution-6-agent-loop): Observe, reason, act. How AI agents work, the ReAct loop, AI agent architecture, and why the execution path is emergent.
- [Agentic Durable Execution: The Loop Escapes the Process](https://www.diagrid.io/blog/agentic-execution-evolution-7-durable-agent-loop): Checkpointing vs durable execution, process supervision vs durable execution, deterministic replay, exactly-once execution, crash recovery, and AI agent execution guarantees.
### Recent Blog Posts
- [AI Agent Orchestration at Scale: How to Size a System for 10,000 Concurrent Workflows](https://www.diagrid.io/blog/ai-agent-orchestration-at-scale)
- [What's New in Dapr 1.19: Faster Workflows, New Data APIs, and a Simpler Control Plane](https://www.diagrid.io/blog/dapr-1-19-release-highlights)
- [Introducing Dapr Durable Execution for Goose: AI Agents That Pick Up Where They Left Off](https://www.diagrid.io/blog/durable-execution-goose-agents)
- [Making n8n Workflows Durable with Dapr](https://www.diagrid.io/blog/durable-n8n-workflows-dapr)
- [AI Agent Frameworks in 2026: The 17 Best Compared](https://www.diagrid.io/blog/ai-agent-frameworks)
- [Dapr Ops Dashboard Is Now Free for the Dapr Community](https://www.diagrid.io/blog/dapr-ops-dashboard-free)
- [Durable Execution Pricing in 2026: How Temporal, Step Functions and Others Bill You](https://www.diagrid.io/blog/durable-execution-pricing-models-2026)
- [8 Questions a CISO Should Ask Every AI Agent Security Vendor](https://www.diagrid.io/blog/ciso-questions-ai-agent-security)
- [The 9 questions bank risk teams ask before an AI agent ships](https://www.diagrid.io/blog/bank-risk-questions-ai-agents)
- [LangGraph in Production: The 12-Point Checklist Before Your Agent Takes Real Traffic](https://www.diagrid.io/blog/langgraph-agent-production-checklist)
- [Run reliable agents and apps without giving up control](https://www.diagrid.io/blog/catalyst-deployment-models)
- [Announcing Durable Execution for Spring AI Agents](https://www.diagrid.io/blog/durable-execution-spring-ai-agents)
- [Why Durable Execution Belongs Below the Spring AI Framework](https://www.diagrid.io/blog/why-durable-execution-belongs-below-spring-ai)
- [Keep Your Agent Framework, Add Durable Execution](https://www.diagrid.io/blog/keep-your-agent-framework)
- [GopherCon 2026: Hidden Flags, Durable Agents, and Bringing the Community Together](https://www.diagrid.io/blog/gophercon-2026-recap)
- [Why Checkpointing Is Not Agentic Durable Execution for Production AI Agents](https://www.diagrid.io/blog/checkpointing-vs-agentic-durable-execution)
- [How Agentic Durable Execution Cuts Rework and Token Spend in Long-Running Agents](https://www.diagrid.io/blog/agentic-durable-execution-cuts-rework-and-token-spend)
- [Patchmageddon Is Coming: What the EU Cyber Resilience Act Means for Dapr Users](https://www.diagrid.io/blog/eu-cyber-resilience-act-dapr)
- [Observability for Non-Deterministic Agent Workflows](https://www.diagrid.io/blog/runtime-native-agent-observability)
- [Durable Execution, Now Built for Agents](https://www.diagrid.io/blog/what-is-agentic-durable-execution)