comparison

Camunda vs Zeebe 2026: Camunda 7 Platform vs Camunda 8 Cloud-Native Engine

Zeebe is the cloud-native BPMN workflow engine that powers Camunda 8, while Camunda 7 is the mature JVM-based platform that preceded it. Both are maintained by Camunda Services GmbH. This 2026 comparison clarifies the architecture differences, feature deltas, migration considerations, and pricing between the two generations.

Overview

Camunda and Zeebe are frequently confused, and the confusion is reasonable because the two names are intertwined in the product history. Zeebe is the cloud-native BPMN workflow engine developed by Camunda, and it is the engine that powers Camunda 8 (the current generation of the platform, launched in April 2022). Camunda 7 is the mature JVM-based platform that preceded it, built around an embedded engine pattern. Both are maintained by Camunda Services GmbH (founded 2008, headquartered in Berlin).

As of April 2026, Camunda 7 is in long-term support (LTS), with maintenance committed through at least 2027. Camunda 8 is the actively developed platform, delivered as a SaaS service and as a self-managed distribution. Teams evaluating Camunda today are effectively choosing between Camunda 7 and Camunda 8 (Zeebe-powered).

Summary Table

Feature Camunda 7 Camunda 8 (Zeebe)
Engine architecture Embedded JVM engine, relational database Distributed gRPC-based engine (Zeebe), event-sourced log
Deployment model Embedded in Spring Boot or standalone server Self-managed Kubernetes or Camunda 8 SaaS
Primary language Java Language-agnostic (gRPC clients in Java, Node.js, C#, Go, Python)
BPMN engine Native BPMN 2.0 Native BPMN 2.0 (subset plus cloud-native extensions)
State persistence JDBC-connected relational DB (PostgreSQL, Oracle, MySQL) Event log in RocksDB, replicated via Raft
Scaling model Vertical scaling, clustered via shared database Horizontal scaling via partition rebalancing
Operator UX Cockpit Operate, Tasklist, Optimize (separate web applications)
Process modeler Camunda Modeler (desktop) Camunda Modeler and Web Modeler (SaaS)
Community edition Free, open source (Apache 2.0) Zeebe is source-available (Camunda Platform License); SaaS has free Starter tier
Enterprise pricing Commercial contract, typical six-figure enterprise Camunda 8 Starter ~$10K/year; self-managed from ~$49K/year; scales by process instances
Latest release Camunda 7.22 (LTS, April 2026) Camunda 8.6 (April 2026)

Architecture Differences

Camunda 7: Embedded Engine

The Camunda 7 engine is a Java library that can be embedded directly inside a Spring Boot application or deployed as a standalone JBoss or Tomcat server. Workflow state is persisted via JDBC to a relational database shared across engine instances. Clustering is achieved by pointing multiple engine JVMs at the same database; a pessimistic locking strategy prevents two engines from working on the same token simultaneously.

Advantages of the embedded model:

  • Transactional consistency with application data (the same JDBC transaction can update business rows and advance the workflow state)
  • Synchronous service tasks via direct Java method calls
  • Simple local development with an H2 in-memory database

Constraints:

  • JVM-only, which couples workflow ownership to Java or Kotlin backends
  • Scaling requires a beefy shared database; high-throughput deployments hit database contention
  • Workflow logic and application logic share a deployment lifecycle

Camunda 8: Zeebe Distributed Engine

Zeebe is a standalone broker cluster that replicates workflow state via a Raft-based consensus protocol across partitions. Clients (job workers) connect over gRPC in any supported language and pull tasks when ready. State is persisted to RocksDB, and the event log is streamed to Elasticsearch for the Operate and Optimize UIs.

Advantages of Zeebe:

  • Horizontal scaling by adding partitions; no shared database bottleneck
  • Language-agnostic workers (Java, Node.js, Python, Go, C#, Rust via community client)
  • Event-sourced log enables time-travel debugging in Operate

Trade-offs:

  • Operational complexity of running a Zeebe cluster on Kubernetes (or paying for SaaS)
  • Loss of shared-JDBC transactional consistency — workflow and business data commits are separate
  • Smaller BPMN coverage in early Camunda 8 releases has since caught up by 8.6, but a handful of advanced Camunda 7 patterns still require redesign

Feature Deltas

Most standard BPMN 2.0 patterns work identically in both versions: sequence flows, gateways, user tasks, service tasks, timers, boundary events, and call activities. Key differences as of April 2026:

  • DMN (Decision Model and Notation): Both engines support DMN. Camunda 7 uses FEEL and JUEL; Camunda 8 uses FEEL exclusively.
  • CMMN (Case Management Model and Notation): Supported in Camunda 7, not supported in Camunda 8.
  • Script tasks with embedded Groovy/JavaScript: Standard in Camunda 7; in Camunda 8 restricted to FEEL expressions or external workers.
  • Message correlation: Both support, though Camunda 8 requires explicit correlation keys defined in the model.
  • Multi-tenancy: Camunda 8 SaaS offers native multi-cluster tenancy; Camunda 7 has schema-level multi-tenancy.

Migration Considerations

Teams on Camunda 7 considering migration to Camunda 8 should evaluate:

  • BPMN coverage: Run existing process models through the Camunda Modeler 8 linter to identify unsupported elements
  • Script task usage: If processes rely on embedded scripts, those must be externalized as job workers
  • CMMN processes: Case-management workflows need redesign as BPMN with ad-hoc subprocesses
  • Transactional consistency: Business logic that depended on shared JDBC transactions must be refactored into Saga-style compensations
  • Licensing: Camunda 8 Enterprise is priced by process instances started per year; compare against Camunda 7 CPU-based licensing

Camunda provides a migration guide and tooling, but most migrations require several weeks of rework for non-trivial models.

When to Stay on Camunda 7

  • Existing JVM-embedded deployments with strong transactional requirements
  • Heavy CMMN case-management usage
  • Shared-database operational model that works well at current scale
  • Stable processes where rewriting for Zeebe offers no throughput benefit

When to Move to Camunda 8

  • Greenfield BPMN deployments with expected throughput above a few thousand instances/day
  • Polyglot environments where workers are written in multiple languages
  • Kubernetes-native platform teams seeking horizontal scalability
  • Organizations that prefer the SaaS operational model over self-hosting

Pricing Summary

Camunda 7 Community is free under the Apache 2.0 license. Camunda 7 Enterprise (with Cockpit, Tasklist, and support) is a commercial contract typically negotiated at the six-figure annual level for large enterprises.

Camunda 8 pricing as of April 2026:

  • Camunda 8 SaaS Free Trial: 30 days
  • Camunda 8 SaaS Starter: approximately $10,000/year for a fixed instance quota and standard support
  • Camunda 8 SaaS Professional / Enterprise: custom pricing by process instance volume
  • Camunda 8 Self-Managed: from approximately $49,000/year for production-grade support and enterprise entitlements

Editor's Note: We ran a migration assessment for a European insurance client on Camunda 7.18 in 2025. They had 140 deployed BPMN models across underwriting and claims, 22 of which used CMMN case patterns. The assessment estimated 14 weeks of engineering work to redesign the CMMN workflows as BPMN ad-hoc subprocesses, plus an additional 6 weeks to externalize 60+ script tasks as Node.js job workers. Annual license delta was +$38K for Camunda 8 Self-Managed. They elected to stay on Camunda 7 LTS until 2027 and revisit closer to the end-of-life window. The main honest caveat: Camunda 8 is the strategic direction of the platform, and new features (like the Web Modeler, native Elasticsearch search, and Zeebe performance improvements) will continue to land there first.

Written & reviewed by Rafal Fila · Last updated:

Tools Mentioned

Related Guides

comparison

Keystroke vs n8n in 2026: Agent-Built TypeScript vs the Visual Canvas

Keystroke, launched in July 2026 by Y Combinator W24 company Sprint Labs, is a code-first automation platform where AI coding agents write workflows as TypeScript in the user's repository. n8n, founded in 2019, is the most widely deployed source-available visual workflow platform, with 200,000+ users and a $2.5 billion valuation. This comparison covers the agent-authored versus canvas building models, durable execution, licensing (Elastic License 2.0 vs the Sustainable Use License), verified July 2026 pricing including Keystroke's usage metering, and the maturity gap between a days-old platform and an established ecosystem.

comparison

QuantumBPM vs Camunda 2026: Single-Binary Challenger vs the BPMN Incumbent

QuantumBPM (launched 2026, Coroid s.r.o., Slovakia) packages a BPMN 2.0 runtime and DMN 1.5 decision engine into one Go binary backed by Temporal and PostgreSQL. Camunda (Berlin, founded 2013) is the category incumbent: Camunda 7 (Apache 2.0, in maintenance) and the Zeebe-based Camunda 8 platform. This comparison covers product structure, architecture, DMN TCK conformance with recording dates, deployment, pricing, and vendor maturity, verified July 2026.

case-study

Migrating 23 Make Scenarios to Self-Hosted n8n: a 3-Week Breakdown

Anonymized retrospective of a DTC ecommerce brand migrating 23 Make scenarios to a self-hosted n8n instance over three weeks. Tooling cost dropped from $348/month on Make Teams to roughly $12/month on a Hetzner VPS, but credential and webhook recreation consumed about 40% of total project time.

Related Rankings

Common Questions

Is Keystroke worth it in 2026?

Keystroke earns a provisional 6.5/10 in its July 2026 open alpha: the agent-native TypeScript model is genuinely differentiated and the free Hobby tier makes it safe to trial, but the platform is pre-1.0 (npm at v0.1.98, public repository published July 13, 2026), has no third-party production track record, and its Elastic License 2.0 restricts offering it as a hosted service. Worth trialing for teams that build through coding agents; too young for production-critical workflows.

How much does Keystroke cost in 2026?

Keystroke offers three tiers as of July 2026: Hobby (free forever, with $1/month of included usage credit), Pro ($20/month, including $20/month of usage credit), and Organization (custom pricing with SSO, RBAC, and audit logs). Usage is metered on every tier: $0.01 per agent or workflow run, $0.005 per empty poll, $0.007 per web search, roughly $0.067 per hour of sandbox compute, and a 1.1x markup on AI model calls unless you bring your own API keys.

What is Keystroke?

Keystroke is a code-first workflow automation and AI agent platform, launched July 13, 2026 by Y Combinator-backed Sprint Labs, that positions itself as an n8n alternative built for AI coding agents. Workflows are written as typed TypeScript in the user's own repository, usually by agents such as Claude Code, Cursor, or Codex, and deployed to Keystroke's managed cloud or self-hosted under the source-available Elastic License 2.0.

Is Keystroke a good n8n alternative?

Keystroke is a credible n8n alternative only for teams that build automations through coding agents such as Claude Code or Cursor, and even then as an experiment rather than a replacement as of July 2026: it launched July 13, 2026 and is pre-1.0, while n8n has 200,000+ users, 196,000+ GitHub stars, and a mature self-hosting story. For visual building, non-developer operators, or production-critical workflows, n8n remains the stronger choice.