Non-JSON OpenAPI Responses

Scope: endpoints emitted by api/dev/generate_swagger_specs.py into the generated console, openapi, service, and web specs.

The current Flask-RESTX generator still emits these response entries under application/json. The schema model names below record the actual runtime response intent so contract generation does not treat them as missing annotations.

Binary And Audio

Spec	Method	Path	Runtime response	Schema
console	POST	/apps/{app_id}/text-to-audio	Audio stream, usually audio/mpeg	AudioBinaryResponse
console	POST	/installed-apps/{installed_app_id}/text-to-audio	Audio stream, usually audio/mpeg	AudioBinaryResponse
console	POST	/trial-apps/{app_id}/text-to-audio	Audio stream, usually audio/mpeg	AudioBinaryResponse
web	POST	/text-to-audio	Audio stream, usually audio/mpeg	AudioBinaryResponse
service	POST	/text-to-audio	Audio stream, usually audio/mpeg	AudioBinaryResponse
console	POST	/datasets/{dataset_id}/documents/download-zip	application/zip attachment	BinaryFileResponse
service	POST	/datasets/{dataset_id}/documents/download-zip	application/zip attachment	BinaryFileResponse
service	GET	/files/{file_id}/preview	Original file MIME type, optionally attachment	BinaryFileResponse
console	GET	/workspaces/current/plugin/icon	Plugin asset MIME type	BinaryFileResponse
console	GET	/workspaces/current/plugin/asset	application/octet-stream	BinaryFileResponse
console	GET	/workspaces/current/tool-provider/builtin/{provider}/icon	Tool icon MIME type	BinaryFileResponse
console	GET	/workspaces/current/trigger-provider/{provider}/icon	Trigger icon response	BinaryFileResponse
console	GET	/workspaces/{tenant_id}/model-providers/{provider}/{icon_type}/{lang}	Model provider icon MIME type	BinaryFileResponse

Text File Exports

Spec	Method	Path	Runtime response	Schema
console	GET	/apps/{app_id}/feedbacks/export	text/csv attachment by default; format=json returns JSON attachment	TextFileResponse
console	GET	/workspaces/current/customized-snippets/{snippet_id}/export	application/x-yaml attachment	TextFileResponse

Fixed SSE Streams

Spec	Method	Path	Runtime response	Schema
console	GET	/workflow/{workflow_run_id}/events	text/event-stream	EventStreamResponse
console	GET	/apps/{app_id}/workflows/draft/runs/{run_id}/node-outputs/events	text/event-stream	EventStreamResponse
console	GET	/apps/{app_id}/workflows/published/runs/{run_id}/node-outputs/events	text/event-stream	EventStreamResponse
openapi	POST	/apps/{app_id}/run	text/event-stream	EventStreamResponse
openapi	GET	/apps/{app_id}/tasks/{task_id}/events	text/event-stream	EventStreamResponse
service	GET	/workflow/{task_id}/events	text/event-stream	EventStreamResponse
web	GET	/workflow/{task_id}/events	text/event-stream	EventStreamResponse

Generated Responses With Streaming Variants

These endpoints call helper.compact_generate_response(...). They return JSON in blocking mode and text/event-stream in streaming mode. Some console/debug paths always pass streaming=True.

Spec	Method	Path	Streaming behavior	Schema
console	POST	/apps/{app_id}/completion-messages	response_mode=streaming	GeneratedAppResponse
console	POST	/apps/{app_id}/advanced-chat/workflows/draft/run	Always streaming	GeneratedAppResponse
console	POST	/apps/{app_id}/advanced-chat/workflows/draft/iteration/nodes/{node_id}/run	Always streaming	GeneratedAppResponse
console	POST	/apps/{app_id}/advanced-chat/workflows/draft/loop/nodes/{node_id}/run	Always streaming	GeneratedAppResponse
console	POST	/apps/{app_id}/workflows/draft/run	Always streaming	GeneratedAppResponse
console	POST	/apps/{app_id}/workflows/draft/iteration/nodes/{node_id}/run	Always streaming	GeneratedAppResponse
console	POST	/apps/{app_id}/workflows/draft/loop/nodes/{node_id}/run	Always streaming	GeneratedAppResponse
console	POST	/apps/{app_id}/workflows/draft/trigger/run	Returns waiting JSON until an event arrives; then streams workflow events	GeneratedAppResponse
console	POST	/apps/{app_id}/workflows/draft/trigger/run-all	Returns waiting JSON until an event arrives; then streams workflow events	GeneratedAppResponse
console	GET	/installed-apps/{installed_app_id}/messages/{message_id}/more-like-this	response_mode=streaming	GeneratedAppResponse
console	POST	/installed-apps/{installed_app_id}/completion-messages	response_mode=streaming	GeneratedAppResponse
console	POST	/installed-apps/{installed_app_id}/chat-messages	Always streaming	GeneratedAppResponse
console	POST	/installed-apps/{installed_app_id}/workflows/run	Always streaming	GeneratedAppResponse
console	POST	/trial-apps/{app_id}/chat-messages	Always streaming	GeneratedAppResponse
console	POST	/trial-apps/{app_id}/completion-messages	response_mode=streaming	GeneratedAppResponse
console	POST	/trial-apps/{app_id}/workflows/run	Always streaming	GeneratedAppResponse
console	POST	/snippets/{snippet_id}/workflows/draft/run	Always streaming	GeneratedAppResponse
console	POST	/snippets/{snippet_id}/workflows/draft/iteration/nodes/{node_id}/run	Always streaming	GeneratedAppResponse
console	POST	/snippets/{snippet_id}/workflows/draft/loop/nodes/{node_id}/run	Always streaming	GeneratedAppResponse
console	POST	/rag/pipelines/{pipeline_id}/workflows/draft/run	Always streaming	RagPipelineOpaqueResponse
console	POST	/rag/pipelines/{pipeline_id}/workflows/published/run	response_mode=streaming	RagPipelineOpaqueResponse
console	POST	/rag/pipelines/{pipeline_id}/workflows/draft/iteration/nodes/{node_id}/run	Always streaming	RagPipelineOpaqueResponse
console	POST	/rag/pipelines/{pipeline_id}/workflows/draft/loop/nodes/{node_id}/run	Always streaming	RagPipelineOpaqueResponse
console	POST	/rag/pipelines/{pipeline_id}/workflows/draft/datasource/nodes/{node_id}/run	Always streaming	RagPipelineOpaqueResponse
console	POST	/rag/pipelines/{pipeline_id}/workflows/published/datasource/nodes/{node_id}/run	Always streaming	RagPipelineOpaqueResponse
service	POST	/completion-messages	response_mode=streaming	GeneratedAppResponse
service	POST	/chat-messages	response_mode=streaming; agent apps are streaming-only	GeneratedAppResponse
service	POST	/workflows/run	response_mode=streaming	GeneratedAppResponse
service	POST	/workflows/{workflow_id}/run	response_mode=streaming	GeneratedAppResponse
service	POST	/datasets/{dataset_id}/pipeline/run	response_mode=streaming	GeneratedAppResponse
service	POST	/datasets/{dataset_id}/pipeline/datasource/nodes/{node_id}/run	Always streaming	GeneratedAppResponse
web	POST	/completion-messages	response_mode=streaming	GeneratedAppResponse
web	POST	/chat-messages	response_mode=streaming; agent apps are streaming-only	GeneratedAppResponse
web	GET	/messages/{message_id}/more-like-this	response_mode=streaming	GeneratedAppResponse
web	POST	/workflows/run	Always streaming	GeneratedAppResponse

Redirects Not Included In Generated Contracts

These console endpoints are browser redirect flows with no 2xx success response. They are documented in the backend OpenAPI as 302 responses with RedirectResponse, but openapi-ts.api.config.ts excludes 3xx-only operations from the oRPC contract input because they are not JSON contract operations.

Spec	Method	Path	Runtime response	Schema
console	GET	/mcp/oauth/callback	Redirect to console OAuth result page	RedirectResponse
console	GET	/oauth/authorize/{provider}	Redirect to OAuth authorization URL	RedirectResponse
console	GET	/oauth/data-source/callback/{provider}	Redirect to console with data source OAuth result	RedirectResponse
console	GET	/oauth/login/{provider}	Redirect to OAuth authorization URL	RedirectResponse
console	GET	/oauth/plugin/{provider_id}/datasource/callback	Redirect to console OAuth callback page	RedirectResponse
console	GET	/oauth/plugin/{provider}/tool/callback	Redirect to console tool OAuth result page	RedirectResponse
console	GET	/oauth/plugin/{provider}/trigger/callback	Redirect to console trigger OAuth result page	RedirectResponse