Camunda vs Flowable (2026)

Agents act through the BPMN activities a builder places in an ad hoc sub process, so any service task, connector call or sub process becomes a tool the model can choose. Camunda offers more than 400 prebuilt connectors, including Salesforce, SAP OData and RFC, HubSpot, Slack, Microsoft Teams, GitHub, Twilio, WhatsApp, Google Drive and Sheets, Amazon S3, DynamoDB and Lambda, plus generic REST, SOAP, GraphQL and SQL. Credentials sit in secrets, a connector SDK builds custom connectors, and MCP client connectors let an agent discover and call tools on outside MCP servers.

Sourcedocs.camunda.io/docs/components/connectors/out-of-the-box-connectors/available-connectors-overviewread 2026-10-12

Agents call tools to query data sources, call APIs and take actions, and tools can be attached to a whole agent or to one operation so a capability is only available where it is needed. Around the agents, cases and processes reach other systems through about 30 connectors, among them Salesforce sObjects, SharePoint documents, ABBYY document capture and Apache Camel, plus HTTP tasks, MuleSoft and Power Automate tasks, a service registry for REST services and an OpenAPI generator. External and A2A agents hand work to Salesforce Agentforce, Azure AI Foundry or any agent that speaks A2A.

Sourcedocumentation.flowable.com/latest/reactmodel/agent/concepts/tools/index.htmlread 2026-10-12

Processes run on the Zeebe engine as BPMN, with gateways, timers, events, sub processes and DMN decisions. An AI agent is an addressable loop inside that process. The model picks the next activity from an ad hoc sub process, the engine runs it as a governed step, and the result goes back to the model until the goal is met. Several agents can share one process, agents built in LangGraph, Amazon Bedrock or custom code can be connected and tracked as agent instances, and predictable steps run straight through while agents handle the judgment calls.

Sourcedocs.camunda.io/docs/components/agentic-orchestration/ai-agentsread 2026-10-12

Work runs on BPMN processes, CMMN cases and DMN decisions, and agents are steps inside them through the AI Agent task. The orchestrator agent coordinates several agents and outside APIs within a case. Each time the case changes it reads what the case allows, reasons about the best next step, and takes it or proposes it. Agent calls can run outside the database transaction, so a long model call does not hold up the case, and the case model decides which data each agent sees and where it may act.

Sourcedocumentation.flowable.com/latest/reactmodel/agent/introduction/index.htmlread 2026-10-12

Processes start or move on without anyone pressing a button. Inbound connectors listen on an HTTP webhook, Kafka, RabbitMQ, Amazon SQS and SNS, Amazon EventBridge, Slack, email, Microsoft 365 mail, GitHub and Twilio, and BPMN timer and message events start or resume instances on a schedule or when a message arrives. An MCP start event registers a deployed process as a tool that outside AI clients can call, and each call starts a new instance.

Sourcedocs.camunda.io/docs/components/connectors/out-of-the-box-connectors/available-connectors-overviewread 2026-10-12

Processes and cases start or move on from outside events. Inbound channels in the event registry listen on Kafka, JMS, RabbitMQ, AWS SQS, email, REST and Apache Camel, and an event registry start subscribes to a channel and starts a new instance, or only one per set of correlation values. Timer, message and signal start events cover schedules and messages, and agents other than the orchestrator can also be started directly over the REST API when the agent model allows it.

Sourcedocumentation.flowable.com/latest/reactmodel/bpmn/reference/event-registry-start/index.htmlread 2026-10-12

Agents can draw on company knowledge kept in a vector store. The Vector Database connector splits documents into chunks, embeds them with Amazon Bedrock Titan, Azure OpenAI, Google Vertex AI or OpenAI models, and writes them to Amazon OpenSearch, OpenSearch, Elasticsearch, Azure AI Search or Azure Cosmos DB. At run time it retrieves the most relevant passages by semantic similarity, so an agent can look up policies and procedures, search a large product catalog, reuse answers to past support questions or quote the exact policy behind a decision. The store grows as new documents are embedded.

Sourcedocs.camunda.io/docs/components/agentic-orchestration/rag-ai-agentsread 2026-10-12

A knowledge base in Flowable Design turns policies, product documentation, FAQs and other content into a semantically indexed store. Content comes in from folder paths or uploads, is extracted and split automatically, and is embedded in Elasticsearch 8, an OpenAI vector store or a custom store, with tunable top K, candidate and similarity settings. Knowledge agents query it at run time, and the same knowledge base can serve processes, forms or a standalone search.

Sourcedocumentation.flowable.com/latest/reactmodel/knowledge-base/introduction/index.htmlread 2026-10-12

Each agent keeps a short term conversation inside its process instance, passing the last 20 messages to the model by default while always keeping the system prompt. Those messages can live in a process variable, in Camunda document storage with a time to live set to the process's expected lifetime, in Amazon Bedrock AgentCore Memory, or in a custom store on self managed setups. Recall from one process instance to the next relies on an outside service such as AgentCore Memory, which AWS runs.

Sourcedocs.camunda.io/docs/components/connectors/out-of-the-box-connectors/agentic-ai-aiagent-subprocessread 2026-10-12

An agent's context is the case it works in. The orchestrator agent reasons from the current case data and state and keeps its chat with users inside the case, so what it knows lasts as long as the case does. Other agents receive the input their operation defines on each call. There is no memory store an agent reads from one case to the next, such as what it learned about a customer last time.

Sourcedocumentation.flowable.com/latest/reactmodel/agent/reference/orchestrator-agent/index.htmlread 2026-10-12

People sit in the process as user tasks, and Tasklist shows each assigned task whether the step before it ran automatically or an AI agent produced the result a person now acts on. A user task can be one of an agent's tools, so the agent asks for help or sign off inside its loop. For MCP tools, a gateway in front of the client lets listing and read operations through and routes selected calls to a user task for confirmation before they run. Task work can also reach people in Microsoft Teams, and teams can build their own task apps on the API.

Sourcedocs.camunda.io/docs/components/connectors/out-of-the-box-connectors/agentic-ai-mcp-client-human-in-the-loopread 2026-10-12

The orchestrator agent can be set to suggest instead of act, presenting its next step as an option in the case chat for a person to accept, and high stakes actions always wait for a human. User tasks, including a four eyes check pattern, put sign off wherever the case needs it. Guardrail agents check content before and after each model call, blocking prompt injection or hate speech, redacting phone numbers, emails and card numbers, and rejecting off topic or non compliant answers.

Sourcedocumentation.flowable.com/latest/reactmodel/agent/reference/orchestrator-agent/index.htmlread 2026-10-12

Camunda runs two identity layers, the Orchestration Cluster admin for runtime work and the management identity for Hub and Optimize, and both connect to an outside OpenID Connect provider such as Entra ID for single sign on, with LDAP and Active Directory on the licensed editions. Authorizations, roles, groups, mapping rules and tenants decide who can start, see and change processes, tasks and other resources. The company holds SOC 2 Type II, ISO 27001 and TISAX and is listed with the Cloud Security Alliance on its trust center.

Sourcetrust.camunda.comread 2026-10-12

Flowable holds ISO 27001 and SOC 2, and its cloud supports single sign on through OAuth and SAML, with Entra ID for identity and SSO on dedicated environments. Each agent sees only the data its task needs and can act only where the case allows, and agents exposed over REST answer only to user groups listed on the agent model. The Work MCP server enforces the same permissions and tenant scoping as the Work REST API, so an assistant can do nothing the signed in user could not. Flowable also runs a public trust center.

Sourceflowable.com/product/enterprise-cloudread 2026-10-12

Operate shows each agent instance's state, such as thinking or calling a tool, its usage metrics and its conversation history grouped by loop iteration, including the tools it picked and the results it got back, for native and connected external agents alike. Health indicators and limit warnings flag an agent that keeps calling the same tool or nears its configured limits. Every process and decision instance can be inspected step by step, audit log APIs record user and user task actions, and Optimize builds reports, dashboards and KPI alerts.

Sourcedocs.camunda.io/docs/components/agentic-orchestration/evaluate-agents/monitor-ai-agentsread 2026-10-12

Agents are audited by default at three levels. The basic level records each agent instance's start and end, status, definition, starting user, calling case and task. The low level exchange adds the operation invoked, the content sent to the model, what the user saw and the request and response tokens. For knowledge agents, the sources used and how closely each matched are stored as well, so retrieval can be tuned. Audit records sit in history tables apart from the running agents and are read in Flowable Control.

Sourcedocumentation.flowable.com/latest/reactmodel/agent/concepts/auditing/index.htmlread 2026-10-12

SaaS customers choose the region for each cluster. On Google Cloud the choices include Belgium, London, Iowa, South Carolina, Montréal, Toronto, Singapore and Sydney, and on AWS they include Frankfurt, Paris, Ohio, Oregon, Virginia and Singapore, each paired with a secondary backup region. Enterprise customers can ask for custom regions. Self managed Camunda runs the whole stack on the customer's own infrastructure, and agent model calls can route through a private endpoint inside the customer's network.

Sourcedocs.camunda.io/docs/components/saas/regionsread 2026-10-12

Flowable Cloud runs on Microsoft Azure with EU and US hosting as standard. Cloud Shared is a multi tenant service with logical data separation, while Cloud Dedicated gives each customer a physically separate environment with its own release schedule, dev, test and staging environments, VPN integrations and other regions on request. The same platform installs on premises or in a customer's own cloud with Docker, Helm charts or WAR files.

Sourceflowable.com/product/enterprise-cloudread 2026-10-12

Camunda Marketplace carries model blueprints and connectors from Camunda, partners and the community, installed straight into a Hub project or downloaded for the desktop modeler. AI agent blueprints such as AI Agent Chat Quick Start and AI Agent Chat With Tools run out of the box on the Camunda provided LLM in SaaS. Partner and community connectors sit outside Camunda's commercial support, and administrators of a self managed Hub can switch the Marketplace off.

Sourcedocs.camunda.io/docs/components/hub/workspace/modeler/modeling/camunda-marketplaceread 2026-10-12

Builders start from agent types, utility, knowledge, document, orchestrator, external, A2A, guardrail and evaluator, each a component they configure with their own prompts and operations. Any model in Flowable Design can be marked as a template so teams copy their own processes, cases and agents, and worked examples include a document analysis agent. There is no catalog of ready made agents or case apps to install for a job such as onboarding or claims.

Sourcedocumentation.flowable.com/latest/reactmodel/model-templates/index.htmlread 2026-10-12

The builder picks the provider and the backend for each agent. Anthropic Claude models run directly, through AWS Bedrock or through Microsoft Foundry, OpenAI models run directly, through Foundry or through any OpenAI compatible endpoint, and Google Gemini runs directly or through Google's Enterprise Agent Platform, with custom endpoints for private network setups and a custom provider on self managed installs. SaaS adds a Camunda provided LLM with a free budget for testing, so teams can switch between models such as Claude Sonnet 4.6 before moving to their own account.

Sourcedocs.camunda.io/docs/components/connectors/out-of-the-box-connectors/agentic-ai-aiagent-model-providersread 2026-10-12

Each agent sets its own vendor, model, temperature, endpoint, custom headers and authentication. Supported vendors are OpenAI, Anthropic and Azure OpenAI, and the OpenAI and Anthropic settings also cover compatible endpoints such as a locally hosted model on Ollama or vLLM. Admins can instead define shared agent connections that agents reference by key. On Cloud Shared, agents connect to the supported providers, and custom models on Cloud Dedicated are still to come.

Sourcedocumentation.flowable.com/latest/reactmodel/agent/concepts/model-settings/index.htmlread 2026-10-12

The Orchestration Cluster REST API drives processes, instances, user tasks, decisions, identities and audit logs, with SDKs for Java, TypeScript, Python, C# and PHP and the c8ctl command line tool. The Orchestration Cluster MCP Server opens operational work such as incidents, user tasks and process instances to AI clients under the same authentication and authorization model as the REST API. Older Operate and Tasklist REST APIs remain available for existing integrations.

Sourcedocs.camunda.io/docs/apis-tools/orchestration-cluster-api-mcp/orchestration-cluster-api-mcp-overviewread 2026-10-12

Flowable Work has a REST API for processes, cases, tasks, users and agents, and the open source engines add Java APIs for embedding. The Work MCP server lets outside AI assistants browse process and case definitions, act on running instances and tasks, look up users, groups and master data and invoke actions, returning the same data as the REST API under the same permissions. It is off by default and needs an AI connector license, and a separate Design MCP server works on models.

Sourcedocumentation.flowable.com/latest/ai/ai-work-mcp-server/index.htmlread 2026-10-12

Camunda Process Test runs an agent process as an integration test, keeping the real model in the loop while mocking outside tool calls. Conditional behavior reacts to whichever tools the agent activates instead of one fixed order, and judge assertions send an output and a plain language expectation to a model that rates the match against a set threshold, alongside semantic similarity checks. Tests run through a Spring Boot starter inside the project's own build.

Sourcedocs.camunda.io/docs/components/agentic-orchestration/evaluate-agents/test-ai-agentsread 2026-10-12

Flowable Design runs automated unit tests for every agent type except the orchestrator before an app goes live. A test runs the agent on a given input, asserts on the response text, structured fields, guardrail outcomes or extracted data, and mocks the services the agent calls, and Flowable Inspect runs the test sets in bulk. In production, evaluators judge each answer for correctness, tone, faithfulness or format, through an evaluator agent or an outside service, and record the verdict without holding up the reply.

Sourcedocumentation.flowable.com/latest/reactmodel/agent/concepts/testing/index.htmlread 2026-10-12

Camunda RPA runs Robot Framework scripts on a standalone RPA worker to operate legacy mainframe applications, desktop applications such as ERP systems and websites that expose no API. Each script runs as a step in a process, so screen work on an old system sits in the same flow as API calls, agent steps and user tasks, and scripts are built and tested before they move to production workers.

Sourcedocs.camunda.io/docs/components/rpa/overviewread 2026-10-12

The Flowable RPA Framework client runs bots built with the open source RPA Framework, either robot files or Robocorp actions and tasks, on a host of the customer's choosing and connects to Flowable over the REST API. A bot runs as a step in a case or process, for example pulling data out of a desktop client application, and hands the result back to the case when it finishes. The client installs with pip.

Sourcedocumentation.flowable.com/latest/develop/external-clients/rpaframework/introduction/index.htmlread 2026-10-12