Agents, people, and systems as one process you can prove, govern, and improve
Camunda orchestrates people, AI agents, systems, and devices in end-to-end processes that scale with your business.
When the process is mission-critical, you need more than automation
Camunda is the process orchestration platform for enterprise AI. It connects AI agents, people, and systems into end-to-end processes that are governed, auditable, and provable — across any cloud, any agent framework, and any system your business depends on. When a process spans more tools than any single vendor was built for, Camunda is the layer that holds it together and shows you exactly what happened.
Available as SaaS or Self-Managed, with the same orchestration engine and APIs for both. Vendor-agnostic by design: any agent framework, any LLM, any cloud — without rewriting when the stack changes. Trusted in production by 700+ enterprises worldwide, including 9 of the top 10 US banks.
Key takeaways
- Coordinates AI agents, humans, and systems across end-to-end business processes in one governed model.
- Built for organizations where a single process spans many systems — none of them designed to talk to each other.
- Goes further than agent frameworks and iPaaS by owning the complete end-to-end flow and proving what happened at every step.
- Trusted for production AI: SOC 2 Type II, ISO/IEC 27001, GDPR-compliant, 99.9% SLA.
The platform
Agents, models, and systems of record will all evolve. Camunda sits in the middle, so what is above and below can swap without rewriting your processes. One end-to-end flow, governed in one place, portable across any cloud.
Why enterprises standardize on Camunda
Composable, never locked in
Orchestrate any agent framework, any LLM, any cloud. Process logic stays portable as standard BPMN, so what you build today still runs when the stack changes.
Learn moreFuture-proof by design
Swap models, change infrastructure, adopt new frameworks. Your orchestration investment holds its value while everything around it churns.
Learn moreGovernance you don’t have to custom-build
Controls baked in: role-based access control, policy enforcement, human-in-the-loop, audit trails, tool allow and deny lists, error handling, and versioned rollouts.
Learn moreScale you can stake your reputation on
Zeebe is built on event streaming with a peer-to-peer, no-central-database architecture, so you can add brokers to handle more load without bottlenecks. Intuit and Barclays run mission-critical work on it.
Learn moreThe control layer your AI strategy is missing
Where agent frameworks reason and integration platforms move data, Camunda decides when and how each gets used, and proves what happened.
Underneath is the Zeebe orchestration engine: distributed, peer-to-peer, with no central database. On top is BPMN (Business Process Model and Notation), the open standard for executable process diagrams that business and engineering can both read. The process model that ships to production is the model your compliance officer audits.
This is agentic orchestration: deterministic control at the macro level, adaptive reasoning at the micro level, in one platform. Agents get autonomy within boundaries you can see, adjust, and audit.
What Camunda replaces, and what it complements
Most categories adjacent to Camunda do one thing well. Camunda’s job is to coordinate them, not duplicate them.
Replaces
Legacy BPM suites
Centralized, low-code-first, gatekept by IT. Camunda is the developer-friendly alternative with the same business-readable process models and none of the vendor lock-in.
Homegrown process automation tooling
Bespoke state machines, custom retries, homegrown monitoring. Camunda gives you those primitives by default. Your team builds business logic, not plumbing.
Per-team workflow engines
Different teams running different orchestrators with no shared visibility. Camunda lets teams stay autonomous without fragmenting capability.
Complements
Agent frameworks
They reason inside a single step. Camunda governs the journey: when to call the agent, how to validate output, when to escalate to a human, what audit trail to write.
Integration platforms (iPaaS)
They move data point-to-point. Camunda orchestrates business outcomes. The difference between an integration and a process.
Process mining
Process mining shows what’s happening across systems. Camunda runs what should happen. Mining identifies the process; Camunda executes it.
How you build processes in Camunda
From discovery to continuous improvement — one platform, the same model throughout.
Discover
Understand what you have before you build.
ProcessOS analyzes existing systems, operational data, and documentation to surface how processes actually work — not how they were designed to work. It identifies where automation stops, where exceptions pile up, and which processes have the highest return on AI investment before a line of BPMN is drawn.
- Process mining from operational data and logs
- As-is process documentation, generated automatically
- Bottleneck and exception hot-spot analysis
- AI opportunity identification and prioritization
- Re-engineering brief for the build phase
Tools in this phase
ProcessOS
AI process intelligence
Discovers how processes actually run, identifies automation opportunities, and generates improvement proposals from production data.
Learn about ProcessOS →One platform, every capability included
Every component shares the same BPMN process model. What business reviews is what engineering deploys and what the operations team monitors. One artifact from design to production.
Modeler
Where business and IT meet on one process model. The file you ship is the file business reviewed and that runs in production.
Learn moreConnectors
Where Camunda meets the rest of your stack. Pre-built connectors for ERP, CRM, ITSM, and major LLM providers, over MCP, A2A, and REST.
Learn moreZeebe
The execution backbone your processes can outlast. Retries, compensation, timeouts, and escalations are handled at the framework level — not coded into every process by hand. When a step fails, the process recovers automatically.
Learn moreOperate
The control room for every running instance. Live execution traces, incident resolution, and the immutable audit trail your regulators ask for.
Learn moreTasklist
Human work as a first-class orchestration step. Assignment rules, escalations, customizable forms — never an afterthought.
Learn moreOptimize
The proof that things are getting better. Throughput, cycle time, agent quality, token cost, and SLA, all anchored to the BPMN process model.
Learn moreConsole
Cluster and project management for Camunda. Create environments, set resource limits, manage access, and monitor usage across your organization.
Console docsIdentity
RBAC, SSO, and SCIM for the full platform. Define roles, manage service accounts, and integrate with your existing identity provider.
Identity docsBuilt for the load that breaks other engines
First American
Decisions made automatically in a single quarter. 87,853 orders, 105,000 unique products, and 58M+ flow node instances in the same period.
Read the story →T-Mobile Austria
Process instances running daily, across ~200 process definitions integrated with 40+ backend systems.
Read case study →The success factors for us, in my opinion, are full control with cloud-native scalability, cross-region active-active architecture, and developer-friendly transformation, enabling us to deliver our features faster to clients.
Murthy Ganti
Director of Software Development, Platforms & AI Accelerator, RBC Clear
Two delivery models for how your business operates
Choose the deployment that matches how your organization operates, and change direction whenever your business needs shift.
Camunda SaaS
- Clouds
- AWS and Google Cloud
- Data residency
- Region-locked clusters; isolated tenant data
- Encryption
- TLS 1.2+ in transit; AES-256 at rest, KMS-managed
- SSO
- SAML 2.0, OIDC, SCIM 2.0
- Audit retention
- 12 months default; extensible by tier
- Attestations
- SOC 2 Type II, ISO/IEC 27001, GDPR
- Best for
- Faster start, no infrastructure operations to manage
Camunda Self-Managed
- Runtime
- Kubernetes (Helm) on AWS, Azure, Google Cloud, OpenShift, on-prem
- Network
- Private VPC; air-gapped supported
- Encryption
- Customer-managed keys (BYOK) supported
- SSO
- SAML 2.0, OIDC, SCIM 2.0; bring your own identity provider
- Audit retention
- Configurable; export to your SIEM
- Compliance
- Inherit Camunda hardening; certify your environment
- Best for
- Strict residency, regulated workloads, custom topologies
Compliant, secure, and fortified for the enterprise
Camunda is a member of the Cloud Security Alliance and maintains SOC 2 Type II compliance, TISAX, and ISO/IEC 27001 certification. Read about our complete list of certifications and security practices in our Trust Center.
Visit the Trust Center
Frequently asked questions
What is agentic orchestration?
The coordination of AI agents, humans, and systems within governed boundaries: deterministic workflow logic plus adaptive agent reasoning in a single platform. Camunda implements it via Agentic BPMN, which orchestrates both the end-to-end journey and each agent’s tool access.
How does Camunda achieve high availability?
Active-active broker replication across availability zones, automatic leader election, and synchronous replication for near-zero RPO. SaaS production tiers carry a 99.9% SLA with sub-minute RTO.
Is Camunda open source? What’s the license?
Core engine components are source-available; enterprise features ship under a commercial license. Either way, your process logic stays portable as standard BPMN, so you own what you build. Licensing details →
How do migration and upgrades work?
Process versioning is first-class: deploy a new version while older instances finish, migrate in-flight instances via Operate, and run canary rollouts against production traffic. SaaS engine upgrades are zero-downtime.
Which agent frameworks and LLMs does Camunda support?
Any of them. Camunda is agent-agnostic by design: orchestrate any agent framework or custom build, calling any LLM provider via MCP, A2A, or REST. Swap models or frameworks without touching process logic.
Why not build this ourselves on existing infrastructure?
You can. For one process. By process three, your team is maintaining its own monitoring UI, audit trail, retry semantics, version migration, and governance framework. That’s undifferentiated infrastructure every orchestrator eventually rebuilds. Camunda provides those primitives by default, so your team builds business logic, not orchestration plumbing.
Does building on Camunda create reusable assets, or does each project start from scratch?
Every process you build becomes a shared, executable artifact the rest of the organisation can reuse. When the next team needs the same integration, decision logic, or sub-process, they reference what’s already running in production — they don’t rebuild it. The more processes you run on Camunda, the faster new ones ship.