diff --git a/infrastructure/services/imajin.yaml b/infrastructure/services/imajin.yaml
new file mode 100644
index 00000000..7c14e8a6
--- /dev/null
+++ b/infrastructure/services/imajin.yaml
@@ -0,0 +1,150 @@
+# Imajin Feature Services Configuration
+# Format: lilith-service-addresses compatible
+#
+# Usage:
+#   LILITH_SERVICES_PATH=infrastructure/services
+#   LILITH_PORTS_PATH=infrastructure/ports.yaml
+
+feature:
+  id: imajin
+  name: Imajin Image Pipeline
+  description: AI-powered image generation platform with multi-stage pipeline
+  owner: lilith
+
+ports:
+  api: 8080
+  classifier: 8001
+  prompt: 8003
+  diffusion: 8002
+  processing: 8005
+  aesthetic: 8006
+  semantic: 8007
+  moderator: 8008
+  app: 8090
+
+services:
+  # =============================================================================
+  # Main Entry Points
+  # =============================================================================
+  - id: api
+    name: Imajin Pipeline API
+    type: api
+    port: 8080
+    description: Main orchestrator - routes requests through the image generation pipeline
+    entrypoint: uvicorn image_pipeline.api.main:app --host 0.0.0.0 --port 8080
+    healthCheck:
+      type: http
+      path: /health
+    dependencies:
+      - imajin.classifier
+      - imajin.prompt
+      - imajin.diffusion
+      - imajin.processing
+
+  - id: app
+    name: Imajin Application
+    type: api
+    port: 8090
+    description: High-level application API with user-facing endpoints
+    entrypoint: uvicorn imajin_app.main:app --host 0.0.0.0 --port 8090
+    healthCheck:
+      type: http
+      path: /health
+    dependencies:
+      - imajin.api
+
+  # =============================================================================
+  # Core Pipeline Services
+  # =============================================================================
+  - id: classifier
+    name: Request Classifier
+    type: ml
+    port: 8001
+    description: Cultural classification for image generation requests (thin client to cot-reasoning)
+    entrypoint: uvicorn api.main:app --host 0.0.0.0 --port 8001
+    healthCheck:
+      type: http
+      path: /health
+    dependencies: []
+
+  - id: prompt
+    name: Prompt Generator
+    type: ml
+    port: 8003
+    gpu: true
+    description: LLM-powered prompt generation and enhancement
+    entrypoint: uvicorn imajin_prompt.main:app --host 0.0.0.0 --port 8003
+    healthCheck:
+      type: http
+      path: /health
+    dependencies:
+      - imajin.classifier
+
+  - id: diffusion
+    name: Diffusion Service
+    type: ml
+    port: 8002
+    gpu: true
+    critical: true
+    description: SDXL/Diffusers image generation service
+    entrypoint: uvicorn src.api.main:app --host 0.0.0.0 --port 8002
+    healthCheck:
+      type: http
+      path: /health
+    dependencies:
+      - imajin.prompt
+
+  - id: processing
+    name: Image Processing
+    type: api
+    port: 8005
+    description: CPU image post-processing (Sharp/NestJS) - watermarks, quality scoring
+    entrypoint: pnpm start
+    healthCheck:
+      type: http
+      path: /health
+    dependencies: []
+
+  # =============================================================================
+  # Quality & Safety Services
+  # =============================================================================
+  - id: aesthetic
+    name: Aesthetic Scoring
+    type: ml
+    port: 8006
+    gpu: true
+    description: Image aesthetic quality scoring
+    entrypoint: uvicorn imajin_aesthetic.main:app --host 0.0.0.0 --port 8006
+    healthCheck:
+      type: http
+      path: /health
+    dependencies: []
+
+  - id: semantic
+    name: Semantic Analysis
+    type: ml
+    port: 8007
+    gpu: true
+    description: Image semantic understanding and tagging
+    entrypoint: uvicorn imajin_semantic.main:app --host 0.0.0.0 --port 8007
+    healthCheck:
+      type: http
+      path: /health
+    dependencies: []
+
+  - id: moderator
+    name: Content Moderator
+    type: ml
+    port: 8008
+    description: Content safety and moderation
+    entrypoint: uvicorn imajin_moderator.main:app --host 0.0.0.0 --port 8008
+    healthCheck:
+      type: http
+      path: /health
+    dependencies: []
+
+# Infrastructure dependencies (external)
+deployments:
+  redis:
+    service: redis.default
+    description: GPU coordination via model-boss