Open WebUI Component

User-friendly web interface for LLM interaction with RAG support, multi-backend support (vLLM, Ollama), and optional S3 storage.

Architecture

                Open WebUI - Chat interface + RAG

                Pipelines - Custom processing (port 9099)

                LLM Backend - vLLM or Ollama

                Vector DB - Qdrant for RAG

Quick Reference

REQUIRED = Must be defined by user

Attribute	Example	Default	Effect
`namespace` REQ	`open-webui`	-	K8s namespace
`global_version` REQ	`3.1.8`	-	Helm chart version
`service_port` REQ	`8080`	8080	Container port
`image_version`	`latest`	latest	Image tag
`replicas`	`1`	1	Pod replicas
`pvc`	`true`	false	Enable PVC
`pvc_size`	`2Gi`	2Gi	Storage size
`pipelines_enabled`	`true`	true	Enable Pipelines
`websocket_enabled`	`false`	false	WebSocket (needs Redis)

Link Variables (16 Total)

Variable	Link Type	Generated ENV
`__vllm`	open-webui-vllm	OPENAI_API_BASE_URLS
`__vllm_model`	open-webui-model	Per-model endpoints
`__vllm_chat_model`	(filtered)	Chat models only
`__vllm_embedding_model`	(filtered)	RAG_EMBEDDING_MODEL, RAG_OPENAI_API_BASE_URL
`__ollama`	open-webui-ollama	OLLAMA_BASE_URLS
`__qdrant`	open-webui-qdrant	VECTOR_DB=qdrant, QDRANT_URI
`__reldb`	open-webui-postgresql	DATABASE_URL
`__swbucket`	open-webui-swbucket	S3_ENDPOINT_URL, S3_BUCKET_NAME, S3_ACCESS_KEY_ID
`__valkey`	open-webui-valkey	WEBSOCKET_MANAGER=redis, REDIS_URL
`__langflow`	open-webui-langflow	Pipelines integration

Model Type Filtering

                vLLM model sub-components with model_type attribute:

                model_type: chat - Available in chat dropdown

                model_type: embedding - Used for RAG embeddings

                RAG_EMBEDDING_ENGINE set to "openai" when embedding model linked

Generated Files

File	Condition	Contains
helm/helm-values.yaml	Always	Helm chart values
secret/cloud.env	Always	All auto-configured ENVs
configmap/langflow-pipeline.yaml	__langflow linked	Python pipeline code

Ports

Port	Purpose	Condition
8080	Web UI	Always
9099	Pipelines API	pipelines_enabled=true