vision-squeezer 0.1.9

LLM-native image optimization middleware & MCP server. Reduces vision model token consumption by snapping to tile boundaries.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

name: vision-squeezer
displayName: VisionSqueezer
description: LLM-native image optimization middleware. Reduces vision model token consumption by snapping images to tile boundaries, removing padding, and re-encoding. Supports Claude, GPT-4o, GPT-5, and Gemini.
license: Elastic-2.0
homepage: https://github.com/eralpozcan/vision-squeezer
repository: https://github.com/eralpozcan/vision-squeezer
icon: https://raw.githubusercontent.com/eralpozcan/vision-squeezer/main/assets/favicon.png

tools:
  - name: optimize_image
    description: Resize and optimize an image for LLM vision APIs. Snaps dimensions to tile boundaries, removes padding, and re-encodes to minimize token consumption.
  - name: sandbox_execute
    description: Apply atomic image operations (crop, grayscale, binarize, resize, contrast, brightness) locally before sending to an LLM.
  - name: get_savings_stats
    description: Retrieve cumulative token and file size savings tracked across all optimizations.

startCommand:
  type: stdio
  command: npx
  args:
    - -y
    - vision-squeezer