$ loading_
Enable non-vision agents to describe images, run OCR, and extract structured data.
The available material is sparse, but the tool appears to send images to an OpenAI-compatible vision model for description, OCR, or structured extraction. It is open source and does not declare a fixed remote endpoint or required secret, so it leans toward caution rather than high risk, but confidence is limited due to missing documentation, low adoption, and unknown maintenance.
The material explicitly states there are no required secrets or environment variables, and there is no evidence that it directly requests sensitive local credentials. However, because its function depends on an “OpenAI-compatible vision model,” real deployments may still require model access credentials configured on the host side, which is undocumented here.
Although no fixed remote host is declared, the description indicates it calls an OpenAI-compatible vision model to process images, which typically means image content may be sent to an external model service. Because the README is missing, the actual egress destination, transmission scope, and whether fully local models are supported cannot be confirmed.
The system flags it as executes-code, and as a stdio MCP server it will run as a local process and handle agent requests. This is a normal capability for MCP tools; the available material does not show requests for unusual system privileges or clearly unrelated high-risk operations.
By function, it at least needs access to user-provided image data to perform description, OCR, or structured extraction. The material does not specify read/write scope, whether it only reads supplied images, or whether it caches outputs, so there is a normal data-access concern around input images and derived text, but no clear sign of overbroad access.
A positive factor is that the source code is publicly available for review. However, it comes from a third-party registry, has 0 stars, no declared license, unknown maintenance status, and no README, which reduces auditability and supply-chain confidence. This alone is not enough for a high-risk rating, but code and dependency review in an isolated environment is advisable.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "read-image-mcp" yet — see the docs or source repo.
Read this receipt image, extract the merchant name, date, total amount, and tax, and return them as JSON; use null for any missing fields.
A structured JSON object with key receipt fields for reimbursement or record entry.
Run OCR on this image and output all Chinese and English text in natural reading order, preserving paragraph structure.
Clean extracted text from the image, ready to copy into a document.
Describe the main content of this chart, extract the title, axis meanings, key trends, and anomalies, and present them as bullet points.
A concise summary of the chart's key information for quick understanding of the visualized data.
Read local images and pass them to LLMs for vision analysis.
Analyze local or remote images with vision LLMs and generate descriptions.
Generate text descriptions from images with fast fallback vision model support.
Enable non-vision AI clients to analyze images with local Ollama vision models.
Let reasoning models inspect images through description-only vision integration.
Extract image color, texture, and shape features for precise comparison and quality assessment.