Skip to content

AI Agent: New connector types and provider config restructure #7213

@maff

Description

@maff

Is your feature request related to a problem? Please describe.

The current AI Agent connector type names (io.camunda.agenticai:aiagent-job-worker:1) and the legacy provider configuration shape (separate openai / azureOpenAi / openaiCompatible entries; googleVertexAi without a backend toggle; Anthropic locked to direct) carry historical baggage that blocks a clean addition of features like the OpenAI Responses API, Claude on AWS Bedrock through the Anthropic SDK platform-backend modules, and a user-supplied custom provider SPI.

Describe the solution you'd like

Introduce new "AI Agent Task" and "AI Agent Sub-process" connector types and element templates with a clean ProviderConfiguration shape (consistent backend field across providers, apiFamily for OpenAI). Use the new shape to add the features that didn't fit the legacy structure: OpenAI Responses API, Anthropic on Bedrock / Vertex / Foundry via the SDK's platform-backend modules, and a custom provider SPI for user-supplied ChatModelApiFactory implementations. The old connector types stay around as deprecated input-rewrite shims that delegate to the new internals.

Describe alternatives you've considered

See the parent epic.

Metadata

Metadata

Assignees

No one assigned

    Type

    Urgency

    None yet

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions