> For the complete documentation index, see [llms.txt](https://docs.cortecs.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.cortecs.ai/integration-examples/workplace-apps/camunda.md).

# Camunda

Use Cortecs models in Camunda workflows

[Camunda](https://docs.camunda.io/) orchestrates business processes. Connect its AI Agent Task connector to Cortecs through an OpenAI-compatible endpoint.

{% hint style="info" %}
Before you begin, create a Cortecs API key by following the [Member quickstart](/member-quickstart.md) and choose a tool-calling model from the [model catalog](https://cortecs.ai/serverlessModels).
{% endhint %}

## 1. Store your API key

In Camunda SaaS, open your cluster's **Connector secrets** tab and create a secret named `CORTECS_API_KEY` containing your Cortecs API key. For Self-Managed, see [connector secret configuration](https://docs.camunda.io/docs/self-managed/components/connectors/connectors-configuration/#secrets).

## 2. Connect to Cortecs

In Modeler, add an **AI Agent Task** connector to a process and configure:

| Field           | Value                                                                                  |
| --------------- | -------------------------------------------------------------------------------------- |
| Provider        | **OpenAI-compatible**                                                                  |
| API endpoint    | `https://api.cortecs.ai/v1`                                                            |
| API key         | `{{secrets.CORTECS_API_KEY}}`                                                          |
| Model           | The exact model ID from the model catalog                                              |
| System prompt   | `You summarize business requests in one sentence.`                                     |
| User prompt     | `Summarize: A customer needs to change the delivery address before their order ships.` |
| Agent context   | `=agent.context`                                                                       |
| Response Format | **Text**                                                                               |
| Result variable | `agent`                                                                                |

For this first request, leave **Tools** empty and keep the default memory storage settings.

## 3. Run the process

Connect a start event to the task and the task to an end event. Deploy to the cluster containing your secret and start an instance. In Operate, inspect `agent.responseText` for the summary.

For tool use and more configuration options, see the [Camunda AI Agent Task guide](https://docs.camunda.io/docs/components/connectors/out-of-the-box-connectors/agentic-ai-aiagent-task/).