- Home
- Skills
- Image & Video
- Image Upscaler :
More screenshots
Works with the AI Agents you already use
Image Upscaler :
Three enhancement tiers, 4K enlargement, and interactive before/after comparison.
$5
Image Upscaler :
Example session with this skill installed
Use Image Upscaler on the attached photo of my two cats. Show me Light Cleanup, Standard Enhancement, and Detail Reconstruction so I can compare the different strengths.
Keep the cats recognizable and retain their pose, framing, gray and white coat colors, and the blanket pattern as closely as the model allows. Focus on fur, whiskers, eye clarity, and fabric texture without adding objects or changing the scene.
Provide a before/after comparison and export the Standard Enhancement version to a 3840px long edge. Resampling is allowed. Report the actual input and output dimensions and label the delivery method.
- Read your context and instructions
- Compiled the image upscaler
Two cats, three enhancement strengths
Illustrative walkthrough using the actual cat images shown in this listing's screenshots. The original is a soft photo of two cats resting together on a patterned blanket. The three generated versions let you compare enhancement strength before choosing a result.
| Version | Enhancement approach | What to inspect |
|---|---|---|
| Light Cleanup | Gentle cleanup and restrained detail enhancement. | Fur edges, whiskers, and whether the softer appearance suits the photo. |
| Standard Enhancement | Stronger clarity and texture enhancement. | Eye detail, coat texture, and the blanket pattern. |
| Detail Reconstruction | More assertive reconstruction of fine texture. | Plausible added detail and any changes to the cats' features or fabric. |
Delivered example
- Three generated cat-image variants, shown in the gallery in the order above.
- A before/after viewer with a draggable Compare divider, a side-by-side view, Fit and 100% inspection, and a Download image control.
- Enhancement candidates shown in the gallery: original 1024 × 769; each generated result 1447 × 1087.
- Standard Enhancement 4K export: 3840 × 2885, resampled with Lanczos from 1447 × 1087. The input file is retained; this is not native model-generated 4K. The gallery screenshots show the enhancement candidates before this export.
- A small aspect-ratio difference is disclosed in the viewer; edges may not align exactly.
Choose your result
Start with Light Cleanup for a restrained appearance, then compare Standard Enhancement for more definition. Inspect Detail Reconstruction closely before choosing it: reconstructed fur or fabric texture is a plausible AI interpretation, not verified recovery of the original detail.
Use the divider to inspect the cats' eyes, noses, whiskers, and blanket pattern. Check at 100% and download the version you prefer. Exact identity, geometry, and pattern preservation are not guaranteed.
Connects securely to your tools. The creator never sees your data.
About this skill
Enhance images with Light Cleanup, Standard Enhancement, or Detail Reconstruction. Choose how much detail may change, with content presets for portraits, products, illustrations, and gradients to guide preservation of composition, identity, text, and visual style.
Export images to 4K with a 3840px long edge while keeping their aspect ratio. Use native model output when available, or enlarge a smaller result with the included Lanczos resampling helper. Resize-only requests can enlarge an existing image without AI redrawing. The workflow keeps original files and records the actual delivery method.
The skill includes a portable before/after viewer with Compare, side-by-side, Fit, 100% inspection, and download of the supplied result without further processing. Python helpers check dimensions, aspect ratio, hashes, and native-output or export-record consistency.
Gallery order: Light Cleanup, Standard Enhancement, Detail Reconstruction. The screenshots show the interactive before/after viewer using actual generated examples.
Resampling increases pixel dimensions; it does not recover verified missing detail. Native generation and preservation quality depend on the connected image provider and tool. Explicit native-only requests remain subject to native-output verification. Detail Reconstruction can synthesize plausible fine texture in approved noncritical areas.
Designed for Codex with an image-generation or editing tool. The local 4K exporter, verification, and comparison helpers require Python 3 and Pillow.
4K export increases pixel dimensions using Lanczos when native output is unavailable; it does not recover verified original detail. The 3840px long-edge export preserves its input ratio with nearest-pixel rounding, so it is 3840 × 2160 only for 16:9 images. Model enhancement may change identity, text, geometry, or aspect ratio; export does not repair those changes. Native-only requests depend on model capability. The verifier checks files and records but cannot authenticate native origin independently. Python 3 and Pillow are required.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 month ago
Requires
Permissions
Reads the user's source image and generated result; writes generation records, delivery copies, and local comparison HTML. Runs the bundled Python helpers. A browser is used for local comparison. The selected image-generation provider receives the reference image for enhancement; use the user's authorized provider. No API key is bundled, and no environment-variable access is required by these helpers.