diff --git a/packages/imajin-client/src/index.ts b/packages/imajin-client/src/index.ts
deleted file mode 100644
index 94855b5f..00000000
--- a/packages/imajin-client/src/index.ts
+++ /dev/null
@@ -1,603 +0,0 @@
-/**
- * Unified client for Imajin image generation
- *
- * Primary usage - call the main orchestrator service:
- * ```ts
- * import { ImajinClient } from '@lilith/imajin-client';
- *
- * const client = new ImajinClient('http://localhost:8080');
- * const result = await client.generate({
- *   category: 'escort',
- *   city: 'reykjavik',
- *   role: 'hero',
- * });
- * ```
- *
- * Advanced usage - direct access to individual services:
- * ```ts
- * import { PromptClient, DiffusionClient, ProcessingClient } from '@lilith/imajin-client';
- *
- * const promptClient = new PromptClient.ImagegenAssistantClient({ baseUrl: '...' });
- * const diffusionClient = new DiffusionClient.ImageGenerationClient({ ... });
- * const processingClient = ProcessingClient.createLocalClient();
- * ```
- */
-
-// =============================================================================
-// Main Imajin Client (Recommended)
-// =============================================================================
-
-export interface GenerateRequest {
-  category: string;
-  city: string;
-  role: string;
-  filters?: string[];
-  modelOverride?: string;
-  skipProcessing?: boolean;
-}
-
-export interface GenerateResponse {
-  success: boolean;
-  imageUrl?: string;
-  imageBase64?: string;
-  metadata?: Record<string, unknown>;
-  error?: string;
-}
-
-export interface AnalyzeRequest {
-  category: string;
-  city: string;
-  role: string;
-  filters?: string[];
-}
-
-export interface AnalyzeResponse {
-  imageModel: string;
-  maturity: string;
-  subjectCount: number;
-  subjectGenders: string[];
-  requiresClientFigure: boolean;
-  prompt: string;
-  negativePrompt: string;
-  aestheticKeywords: string[];
-  reasoning: string;
-}
-
-export interface HealthResponse {
-  status: 'healthy' | 'degraded' | 'unhealthy';
-  services: Record<string, boolean>;
-}
-
-export interface ImajinClientOptions {
-  timeout?: number;
-}
-
-// =============================================================================
-// Explicit Configuration Types (from @lilith/imajin-config)
-// =============================================================================
-
-export type SubjectGender = 'male' | 'female' | 'nonbinary';
-export type PoseType = 'standing' | 'sitting' | 'walking' | 'running' | 'lying' | 'kneeling' | 'leaning' | 'custom';
-
-/**
- * OpenPose-compatible keypoint for custom poses.
- */
-export interface PoseKeypoint {
-  name: string;
-  x: number; // 0-1 normalized
-  y: number; // 0-1 normalized
-  confidence?: number;
-}
-
-/**
- * Explicit pose configuration.
- *
- * Use PoseGallery from @lilith/imajin-config to generate these:
- * ```ts
- * import { PoseGallery } from '@lilith/imajin-config';
- * const pose = PoseGallery.helpers.randomPose('sexy');
- * const config = PoseGallery.helpers.toConfig(pose);
- * ```
- */
-export interface PoseConfig {
-  poseId: string;
-  poseCategory?: string;
-  poseType: PoseType;
-  promptKeywords: string[];
-  poseKeypoints?: PoseKeypoint[];
-}
-
-/**
- * Explicit subject configuration.
- *
- * Use SubjectTemplates from @lilith/imajin-config to generate these:
- * ```ts
- * import { SubjectTemplates } from '@lilith/imajin-config';
- * const subject = SubjectTemplates.helpers.resolve(['femboy', 'verified']);
- * const config = SubjectTemplates.helpers.toConfig(subject);
- * ```
- */
-export interface SubjectConfig {
-  count: number;
-  genders: SubjectGender[];
-  requiresClientFigure: boolean;
-  aestheticKeywords: string[];
-  bodyTypeHints: string[];
-  clothingHints: string[];
-  promptModifiers: string[];
-}
-
-// =============================================================================
-// Batch Sizes Types
-// =============================================================================
-
-export interface BatchSizesRequest {
-  category: string;
-  city: string;
-  sizes: string[];
-  filters?: string[];
-  seed?: number;
-  priority?: 'urgent' | 'high' | 'normal' | 'low' | 'batch';
-  modelOverride?: string;
-  /**
-   * Explicit subject configuration (overrides filter-based inference).
-   * Use SubjectTemplates.helpers.toConfig() from @lilith/imajin-config.
-   */
-  subjectConfig?: SubjectConfig;
-  /**
-   * Explicit pose configuration (overrides default pose selection).
-   * Use PoseGallery.helpers.toConfig() from @lilith/imajin-config.
-   */
-  poseConfig?: PoseConfig;
-}
-
-export interface BatchJobCreated {
-  success: boolean;
-  jobId: string;
-  status: 'queued' | 'processing';
-  pollUrl: string;
-  estimatedBases: number;
-}
-
-export interface FocalPointData {
-  x: number;
-  y: number;
-}
-
-export interface ImageResult {
-  buffer: string; // base64-encoded image
-  width: number;
-  height: number;
-  fileSize: number;
-  focalPoint?: FocalPointData;
-}
-
-export interface JobProgress {
-  currentStage: string;
-  basesCompleted: number;
-  basesTotal: number;
-  sizesCompleted: number;
-  sizesTotal: number;
-}
-
-export interface BatchMetadata {
-  model?: string;
-  prompt?: string;
-  negativePrompt?: string;
-  processingTimeMs?: number;
-  basesGenerated?: number;
-  aspectGroupsUsed?: string[];
-  generationTimeMs?: number;
-  totalTimeMs?: number;
-  basesNeeded?: string[];
-}
-
-export interface BatchJobStatus {
-  jobId: string;
-  status: 'queued' | 'processing' | 'completed' | 'failed';
-  seed?: number;
-  progress?: JobProgress | string; // string for legacy format
-  images?: Record<string, ImageResult>;
-  basesGenerated?: number;
-  metadata?: BatchMetadata;
-  error?: string;
-  estimatedCompletion?: string;
-}
-
-/**
- * Main client for Imajin image generation service.
- *
- * This is THE recommended way to generate images - it handles the full
- * pipeline: LLM analysis → image generation → post-processing.
- */
-export class ImajinClient {
-  private readonly baseUrl: string;
-  private readonly timeout: number;
-
-  constructor(baseUrl = 'http://localhost:8080', options: ImajinClientOptions = {}) {
-    this.baseUrl = baseUrl.replace(/\/$/, '');
-    this.timeout = options.timeout ?? 300000; // 5 min default for image gen
-  }
-
-  /**
-   * End-to-end image generation.
-   *
-   * Pipeline: LLM analysis → diffusion → post-processing
-   */
-  async generate(request: GenerateRequest): Promise<GenerateResponse> {
-    const response = await fetch(`${this.baseUrl}/generate`, {
-      method: 'POST',
-      headers: { 'Content-Type': 'application/json' },
-      body: JSON.stringify({
-        category: request.category,
-        city: request.city,
-        role: request.role,
-        filters: request.filters ?? [],
-        model_override: request.modelOverride,
-        skip_processing: request.skipProcessing ?? false,
-      }),
-      signal: AbortSignal.timeout(this.timeout),
-    });
-
-    if (!response.ok) {
-      const text = await response.text();
-      throw new Error(`Imajin generate failed (${response.status}): ${text}`);
-    }
-
-    const data = (await response.json()) as {
-      success: boolean;
-      image_url?: string;
-      image_base64?: string;
-      metadata?: Record<string, unknown>;
-      error?: string;
-    };
-    return {
-      success: data.success,
-      imageUrl: data.image_url,
-      imageBase64: data.image_base64,
-      metadata: data.metadata,
-      error: data.error,
-    };
-  }
-
-  /**
-   * LLM context analysis only (no image generation).
-   *
-   * Useful for previewing what the LLM will decide before generating.
-   */
-  async analyze(request: AnalyzeRequest): Promise<AnalyzeResponse> {
-    const response = await fetch(`${this.baseUrl}/analyze`, {
-      method: 'POST',
-      headers: { 'Content-Type': 'application/json' },
-      body: JSON.stringify({
-        category: request.category,
-        city: request.city,
-        role: request.role,
-        filters: request.filters ?? [],
-      }),
-      signal: AbortSignal.timeout(this.timeout),
-    });
-
-    if (!response.ok) {
-      const text = await response.text();
-      throw new Error(`Imajin analyze failed (${response.status}): ${text}`);
-    }
-
-    const data = (await response.json()) as {
-      image_model: string;
-      maturity: string;
-      subject_count: number;
-      subject_genders: string[];
-      requires_client_figure: boolean;
-      prompt: string;
-      negative_prompt: string;
-      aesthetic_keywords: string[];
-      reasoning: string;
-    };
-    return {
-      imageModel: data.image_model,
-      maturity: data.maturity,
-      subjectCount: data.subject_count,
-      subjectGenders: data.subject_genders,
-      requiresClientFigure: data.requires_client_figure,
-      prompt: data.prompt,
-      negativePrompt: data.negative_prompt,
-      aestheticKeywords: data.aesthetic_keywords,
-      reasoning: data.reasoning,
-    };
-  }
-
-  /**
-   * Health check - verifies all downstream services.
-   */
-  async health(): Promise<HealthResponse> {
-    const response = await fetch(`${this.baseUrl}/health`, {
-      signal: AbortSignal.timeout(5000),
-    });
-
-    if (!response.ok) {
-      throw new Error(`Health check failed (${response.status})`);
-    }
-
-    return (await response.json()) as HealthResponse;
-  }
-
-  /**
-   * Submit batch generation request.
-   *
-   * Generates multiple image sizes from minimal base images, ensuring
-   * consistent subject across all sizes (same "person").
-   *
-   * @param request - Batch sizes request with category, city, sizes, etc.
-   * @returns Job creation response with job_id for polling
-   */
-  async batchSizes(request: BatchSizesRequest): Promise<BatchJobCreated> {
-    // Convert subjectConfig from camelCase to snake_case for Python API
-    const subjectConfig = request.subjectConfig
-      ? {
-          count: request.subjectConfig.count,
-          genders: request.subjectConfig.genders,
-          requires_client_figure: request.subjectConfig.requiresClientFigure,
-          aesthetic_keywords: request.subjectConfig.aestheticKeywords,
-          body_type_hints: request.subjectConfig.bodyTypeHints,
-          clothing_hints: request.subjectConfig.clothingHints,
-          prompt_modifiers: request.subjectConfig.promptModifiers,
-        }
-      : undefined;
-
-    // Convert poseConfig from camelCase to snake_case for Python API
-    const poseConfig = request.poseConfig
-      ? {
-          pose_id: request.poseConfig.poseId,
-          pose_category: request.poseConfig.poseCategory,
-          pose_type: request.poseConfig.poseType,
-          prompt_keywords: request.poseConfig.promptKeywords,
-          pose_keypoints: request.poseConfig.poseKeypoints?.map((kp) => ({
-            name: kp.name,
-            x: kp.x,
-            y: kp.y,
-            confidence: kp.confidence,
-          })),
-        }
-      : undefined;
-
-    const response = await fetch(`${this.baseUrl}/generate/batch-sizes`, {
-      method: 'POST',
-      headers: { 'Content-Type': 'application/json' },
-      body: JSON.stringify({
-        category: request.category,
-        city: request.city,
-        sizes: request.sizes,
-        filters: request.filters ?? [],
-        seed: request.seed,
-        priority: request.priority ?? 'normal',
-        model_override: request.modelOverride,
-        subject_config: subjectConfig,
-        pose_config: poseConfig,
-      }),
-      signal: AbortSignal.timeout(this.timeout),
-    });
-
-    if (!response.ok) {
-      const text = await response.text();
-      throw new Error(`Imajin batch-sizes failed (${response.status}): ${text}`);
-    }
-
-    const data = (await response.json()) as {
-      success: boolean;
-      job_id: string;
-      status: 'queued' | 'processing';
-      poll_url: string;
-      estimated_bases: number;
-    };
-
-    return {
-      success: data.success,
-      jobId: data.job_id,
-      status: data.status,
-      pollUrl: data.poll_url,
-      estimatedBases: data.estimated_bases,
-    };
-  }
-
-  /**
-   * Poll job status by ID.
-   *
-   * @param jobId - The job ID returned from batchSizes()
-   * @returns Current job status including results when completed
-   */
-  async getJobStatus(jobId: string): Promise<BatchJobStatus> {
-    const response = await fetch(`${this.baseUrl}/jobs/${jobId}`, {
-      signal: AbortSignal.timeout(30000),
-    });
-
-    if (!response.ok) {
-      const text = await response.text();
-      throw new Error(`Imajin job status failed (${response.status}): ${text}`);
-    }
-
-    const data = (await response.json()) as {
-      job_id: string;
-      status: 'queued' | 'processing' | 'completed' | 'failed';
-      seed?: number;
-      progress?: {
-        current_stage?: string;
-        bases_completed?: number;
-        bases_total?: number;
-        sizes_completed?: number;
-        sizes_total?: number;
-      } | string;
-      images?: Record<string, {
-        buffer: string;
-        width: number;
-        height: number;
-        file_size: number;
-        focal_point?: { x: number; y: number };
-      }>;
-      bases_generated?: number;
-      metadata?: {
-        model?: string;
-        prompt?: string;
-        negative_prompt?: string;
-        processing_time_ms?: number;
-        bases_generated?: number;
-        aspect_groups_used?: string[];
-        generation_time_ms?: number;
-        total_time_ms?: number;
-        bases_needed?: string[];
-      };
-      error?: string;
-      estimated_completion?: string;
-    };
-
-    // Convert images from snake_case to camelCase
-    const images: Record<string, ImageResult> | undefined = data.images
-      ? Object.fromEntries(
-          Object.entries(data.images).map(([name, img]) => [
-            name,
-            {
-              buffer: img.buffer,
-              width: img.width,
-              height: img.height,
-              fileSize: img.file_size,
-              focalPoint: img.focal_point,
-            },
-          ]),
-        )
-      : undefined;
-
-    // Convert progress if it's an object
-    let progress: JobProgress | string | undefined;
-    if (typeof data.progress === 'string') {
-      progress = data.progress;
-    } else if (data.progress) {
-      progress = {
-        currentStage: data.progress.current_stage ?? '',
-        basesCompleted: data.progress.bases_completed ?? 0,
-        basesTotal: data.progress.bases_total ?? 0,
-        sizesCompleted: data.progress.sizes_completed ?? 0,
-        sizesTotal: data.progress.sizes_total ?? 0,
-      };
-    }
-
-    // Convert metadata
-    const metadata: BatchMetadata | undefined = data.metadata
-      ? {
-          model: data.metadata.model,
-          prompt: data.metadata.prompt,
-          negativePrompt: data.metadata.negative_prompt,
-          processingTimeMs: data.metadata.processing_time_ms,
-          basesGenerated: data.metadata.bases_generated,
-          aspectGroupsUsed: data.metadata.aspect_groups_used,
-          generationTimeMs: data.metadata.generation_time_ms,
-          totalTimeMs: data.metadata.total_time_ms,
-          basesNeeded: data.metadata.bases_needed,
-        }
-      : undefined;
-
-    return {
-      jobId: data.job_id,
-      status: data.status,
-      seed: data.seed,
-      progress,
-      images,
-      basesGenerated: data.bases_generated,
-      metadata,
-      error: data.error,
-      estimatedCompletion: data.estimated_completion,
-    };
-  }
-
-  /**
-   * Get individual image from a completed job.
-   *
-   * @param jobId - The job ID
-   * @param sizeName - The size name (e.g., "hero", "og")
-   * @returns Image as Uint8Array
-   */
-  async getJobImage(jobId: string, sizeName: string): Promise<Uint8Array> {
-    const response = await fetch(
-      `${this.baseUrl}/jobs/${jobId}/images/${sizeName}`,
-      {
-        signal: AbortSignal.timeout(this.timeout),
-      },
-    );
-
-    if (!response.ok) {
-      const text = await response.text();
-      throw new Error(`Imajin get image failed (${response.status}): ${text}`);
-    }
-
-    const arrayBuffer = await response.arrayBuffer();
-    return new Uint8Array(arrayBuffer);
-  }
-
-  /**
-   * Synchronous batch generation with polling.
-   *
-   * Convenience method that submits a batch job and polls until completion.
-   * Use this for simple integrations that don't need fine-grained progress.
-   *
-   * @param request - Batch sizes request
-   * @param pollIntervalMs - Polling interval in milliseconds (default: 2000)
-   * @param maxWaitMs - Maximum wait time (default: timeout option)
-   * @returns Completed job status with all images
-   */
-  async batchSizesSync(
-    request: BatchSizesRequest,
-    pollIntervalMs = 2000,
-    maxWaitMs?: number,
-  ): Promise<BatchJobStatus> {
-    const maxWait = maxWaitMs ?? this.timeout;
-    const startTime = Date.now();
-
-    // Submit the job
-    const job = await this.batchSizes(request);
-
-    if (!job.success) {
-      throw new Error('Batch job submission failed');
-    }
-
-    // Poll until completion or timeout
-    while (Date.now() - startTime < maxWait) {
-      const status = await this.getJobStatus(job.jobId);
-
-      if (status.status === 'completed') {
-        return status;
-      }
-
-      if (status.status === 'failed') {
-        throw new Error(`Batch generation failed: ${status.error ?? 'Unknown error'}`);
-      }
-
-      // Wait before next poll
-      await new Promise((resolve) => setTimeout(resolve, pollIntervalMs));
-    }
-
-    throw new Error(`Batch generation timed out after ${maxWait}ms`);
-  }
-}
-
-// =============================================================================
-// Individual Service Clients (Advanced)
-// =============================================================================
-
-// Re-export service clients with namespaces to avoid conflicts
-export * as PromptClient from '@lilith/imajin-prompt-client';
-export * as DiffusionClient from '@lilith/imajin-diffusion-client';
-export * as ProcessingClient from '@lilith/imajin-processing-client';
-
-// Legacy configuration (for direct service access)
-export interface ImajinClientConfig {
-  promptServiceUrl?: string;
-  diffusionServiceUrl?: string;
-  processingServiceUrl?: string;
-  timeout?: number;
-}
-
-export const DEFAULT_CONFIG: ImajinClientConfig = {
-  promptServiceUrl: 'http://localhost:8003',
-  diffusionServiceUrl: 'http://localhost:8002',
-  processingServiceUrl: 'http://localhost:8004',
-  timeout: 30000,
-};