**Version:** 1.2.0 **Updated:** 2026-06-30 **Freshly updated:** v1.2.0 replaces per-image `respira_update_media` loops with a single `respira_update_media_batch` pass (up to ~50 items at once), snapshots the media library with `respira_get_snapshot` before the batch so the whole
Installs just this skill. Get the whole plugin for auto-invocation.
⚡ How it fires
How this skill gets triggered: by you, by Claude, or both.
Fires itselfClaude auto-loads it when your prompt matches the work.
You can call itInvoke it directly when you want it.
Slash command/wordpress-ai-image-optimizer
👁️ Context preview
The summary Claude sees to decide when to auto-load this skill.
**Version:** 1.2.0 **Updated:** 2026-06-30 **Freshly updated:** v1.2.0 replaces per-image `respira_update_media` loops with a single `respira_update_media_batch` pass (up to ~50 items at once), snapshots the media library with `respira_get_snapshot` before the batch so the whole
# WordPress AI Image Optimizer
**Version:** 1.2.0
**Updated:** 2026-06-30
**Freshly updated:** v1.2.0 replaces per-image `respira_update_media` loops with a single `respira_update_media_batch` pass (up to ~50 items at once), snapshots the media library with `respira_get_snapshot` before the batch so the whole run is one-step reversible, runs the audit through `respira_analyze_images`, surfaces optional stock replacements via `respira_search_stock_images` + `respira_sideload_image`, and closes with a `respira_generate_activity_report` summary (e.g. "optimized 47 images, saved 1.2GB").
AI-powered image optimization for WordPress sites. Automatically compresses, converts formats, renames files, and updates all references. Processes images locally in AI code editor, creates optimized versions, and preserves originals for safety.
## What This Skill Does
**Scans:**
- All media library images
- Images in posts and pages content
- Featured images
- Images in page builder elements (Elementor, Divi, Bricks, etc.)
- WooCommerce product images
- Images in custom post types
**Detects:**
- File sizes (actual KB/MB on disk)
- Image dimensions (width x height in pixels)
- Format (JPG, PNG, GIF, WebP, AVIF, SVG)
- Whether width/height attributes are set in HTML
- Whether alt text exists (and quality)
- Whether image is being used (orphaned media)
- Whether responsive image sizes exist (WordPress srcset)
- Compression quality (estimated from file size)
2. Analyzes each image (use `respira_analyze_images` to drive the audit; it returns per-image findings in one call):
- File size and format
- Actual dimensions vs displayed dimensions
- Alt text presence and quality
- Filename SEO-friendliness
- Usage (which pages/posts use this image)
- Responsive image sizes (srcset)
- Lazy loading implementation
3. Detects optimization plugins/CDN:
- Smush, ShortPixel, EWWW, Imagify, Optimole
- Cloudflare, BunnyCDN, other CDNs
4. Scores opportunities by impact:
- File size x page traffic = optimization priority
- Separates quick wins from big wins
5. Generates comprehensive report
**Phase 2: User Approval**
Presents report and asks which optimizations to run:
- Compress oversized images
- Convert JPG/PNG -> WebP
- Resize unnecessarily large originals
- Rename non-SEO filenames
- Add missing alt text
- Regenerate responsive image sizes
User can choose:
- "Optimize everything" (all optimizations)
- "Compress only" (no format changes)
- "Top 10 images" (highest impact first)
- "Homepage images only" (critical pages first)
**Phase 3: AI Processing**
6. Snapshot the media library before any writes with `respira_get_snapshot` so the entire batch is reversible in one step (`respira_restore_snapshot`). Take this once, before the batch — not per image.
7. Downloads images from WordPress to local workspace.
8. Processes images locally:
- Compression
- Format conversion
- Resizing
- SEO-friendly renaming
- AI alt text generation
- Optional: when an image is low quality or simply wrong for the page, find a replacement with `respira_search_stock_images` and pull it in with `respira_sideload_image` instead of re-optimizing the original
9. Uploads optimized versions back to WordPress:
- Creates NEW media entries (original files untouched)
- Applies metadata in one pass with `respira_update_media_batch` rather than one `respira_update_media` call per image. The batch accepts up to ~50 items, so a typical library is updated in a single call (alt text, title, caption, description). Fall back to single `respira_update_media` only for one-off corrections.
- Marks old files as "deprecated - replaced by [new-filename]"
10. Updates content references in duplicates:
- Creates duplicates of affected posts/pages
- Replaces old image URLs with new optimized URLs
- Updates builder and WooCommerce image references
11. Regenerates responsive image sizes and srcset support
**Phase 4: Review & Publish**
12. Provides before/after comparison:
- File size savings
- Format changes
- Dimension changes
- Filename improvements
- Pages affected
13. User reviews duplicates in WordPress admin
14. User approves publishing
15. Old files remain available for safety and rollback
16. Generate a run summary with `respira_generate_activity_report` so the work is auditable in one line (e.g. "optimized 47 images, saved 1.2GB, added alt text to 31"). This reads from the change log Respira already recorded for the batch.
## Honest Disclaimer
This skill downloads images, processes them with AI, and uploads optimized versions. You review all changes before publishing.
**What this skill CANNOT do:**
- Guarantee perfect quality (compression is lossy)
- Fix already broken/corrupted image files
- Optimize CSS background images
- Process images hosted externally outside WordPress uploads
- Improve composition/art direction (technical optimization only)
**What this skill CAN do:**
- Compress images without visible quality loss
- Convert to WebP for significant size reduction
- Resize oversized originals to practical dimensions
- Rename files with SEO-friendly names
- Generate descriptive alt text
- Update references automatically in duplicates
- Preserve originals for safety and rollback
## Safety Model
- Non-destructive: Creates NEW optimized images and keeps old files
- Duplicate-first: Updates content in duplicates, never live pages
- User approval required: Review before publishing
- Rollback ready: Old files preserved and reversible
- No live site impact until explicit approval
## Technical Details
Uses these Respira MCP tools:
**Audit:**
- `respira_analyze_images` (drives the Phase 1 audit; per-image findings in one call)
**Media Operations:**
- `respira_list_media`
- `respira_get_media`
- `respira_upload_media`
- `respira_update_media_batch` (primary metadata writer — up to ~50 items per call)
