Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Image Unblur image preflight MCP check if my 1200x800 PNG fits the upload limits, then plan a Basic pass"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A small local helper for preparing photos before using Image Unblur. The official website performs AI enhancement; this server does not upload, process, or reconstruct image pixels.
The check_image_metadata tool checks a file's format, byte size, and dimensions against the site's published limits: JPG/PNG/WebP, 20 MB, 24 megapixels, and 64–16,384 pixels per side. plan_enhancement suggests a cautious Basic or Pro first pass from a blur description, with reminders to verify faces and text. get_official_links returns the official site, pricing, and contact pages.
For example, before trying to unblur image online, check that a 1200×800 PNG is within the upload limits, then compare any result at full size. The site says Pro can redraw details; a sharper face or letter is not necessarily a more accurate one. Do not use enhancement to claim recovery of intentionally obscured information.
Node.js 20.18.1 or newer:
npm install
npm test
node index.mjsConfigure an MCP client to start node /absolute/path/to/imageunblur-mcp/index.mjs over stdio. No API key or image upload is involved. Limits may change; check the official site before sending a file. MIT license; see LICENSE.
Check only file format, byte size, and dimensions against published Image Unblur limits. No image file is sent.
Inputs: width (required), format (required), height (required), sizeBytes (required).
Return allowlisted official product links.
Inputs: no named parameters.
Choose a cautious first model from blur symptoms. It does not inspect or modify an image.
Inputs: blurType (required), severity (required), containsTextOrFaces (required).
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 3 tools
The three tools target clearly separate preflight stages: returning official resources, checking metadata limits, and planning a model choice. None overlap in action or output, so an agent can easily select the right tool.
All names use consistent snake_case with a verb_noun structure: get_official_links, check_image_metadata, plan_enhancement. The convention is predictable and readable.
Three tools fit the narrow preflight scope, but the surface is very small and one tool is essentially a static link lookup. It is appropriately scoped rather than excessive, though slightly minimal.
For a preflight-only server, the core checks are covered: resource links, metadata validation, and enhancement planning. There is no actual enhancement execution, which is consistent with the stated scope, but agents may need external follow-through via the official links.
Hosted MCP server for AI photo editing: nothing to install, no desktop photo editor needed.
MCP server for Pixapi: check live credit pricing and balance, then generate images and video.
MCP server for generating rough-draft project plans from natural-language prompts.
MCP server exposing the YourImageShare upload API (upload/list/delete) as tools for AI agents.
