Skip to main content
Glama
RunhuaHuang

VisionPower

by RunhuaHuang
OverviewSchemaSourceRelated ServersDiscussionsAdmin
JavaScript
Hybrid

Describe Image

describe_image

Enables text-only models to understand images: reads text, describes scenes, interprets charts, compares images, and returns results as text.

Instructions

Image-understanding tool for non-multimodal (text-only) models. Uses an external vision model to read text (OCR), describe scenes, interpret charts and diagrams, compare images, and return the results as text. Models that can understand images directly must not call this tool; use their own vision capability instead. Accepts image_path, image_url, image_base64, image_ref, or images[]. For faster, more useful answers, ask the specific question you need answered (e.g. "read the error text", "what does this chart show") instead of an open-ended "describe everything".

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
imagesNoOrdered list of images to analyze. Use this for multiple images; do not combine it with top-level image fields.
promptNoSpecific question or instruction about the image(s). Focused questions (e.g. "Read the error message text", "List the visible menu items") are faster and more useful than open-ended "describe everything" prompts. Leave empty for a full description.
image_refNoShort-lived opaque reference created by the VisionPower WebUI Inbox. Use this when a text-only host cannot pass an attachment through to the agent.
image_urlNoPublic http(s) URL of an image; support depends on the configured provider/model. Use image_base64 or image_ref when URL input is unavailable.
image_pathNoAbsolute path to a local raster image file. Use this when the image is available on disk.
image_base64NoBase64-encoded image data without a data: URI prefix.
output_formatNoOutput shape. 'text' (default) returns a free-form description with an untrusted-source banner. 'structured' returns a JSON envelope: when formatValid is true, a single image has {answer, observations, extractedText?, limitations?} and multiple images have images[]; otherwise formatValid is false with formatError and rawResponse.
image_mime_typeNoMIME type for image_base64. If omitted, VisionPower detects it from image bytes.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed5 schema fields changedv3.1.1
    • addedInput schema / properties / image_ref
      Added value: +{
+  "description": "Short-lived opaque reference created by the VisionPower WebUI Inbox. Use this when a text-only host cannot pass an attachment through to the agent.",
+  "pattern": "^vpimg_[A-Za-z0-9_-]{32}$",
+  "type": "string"
+}
    • changedInput schema / properties / image_url / description
      Previous value: -"Public http(s) URL of an image that the configured vision model provider can access."New value: +"Public http(s) URL of an image; support depends on the configured provider/model. Use image_base64 or image_ref when URL input is unavailable."
    • addedInput schema / properties / images / items / properties / image_ref
      Added value: +{
+  "description": "Short-lived opaque reference created by the VisionPower WebUI Inbox. Use this when a text-only host cannot pass an attachment through to the agent.",
+  "pattern": "^vpimg_[A-Za-z0-9_-]{32}$",
+  "type": "string"
+}
    • changedInput schema / properties / images / items / properties / image_url / description
      Previous value: -"Public http(s) URL of an image that the configured vision model provider can access."New value: +"Public http(s) URL of an image; support depends on the configured provider/model. Use image_base64 or image_ref when URL input is unavailable."
    • changedInput schema / properties / prompt / description
      Previous value: -"Specific question or instruction about the image(s). Leave empty for a full description."New value: +"Specific question or instruction about the image(s). Focused questions (e.g. \"Read the error message text\", \"List the visible menu items\") are faster and more useful than open-ended \"describe everything\" prompts. Leave empty for a full description."
  2. Changed1 schema field changedv2.4.2
    • addedInput schema / properties / output_format
      Added value: +{
+  "description": "Output shape. 'text' (default) returns a free-form description with an untrusted-source banner. 'structured' returns a JSON envelope: when formatValid is true, a single image has {answer, observations, extractedText?, limitations?} and multiple images have images[]; otherwise formatValid is false with formatError and rawResponse.",
+  "enum": [
+    "text",
+    "structured"
+  ],
+  "type": "string"
+}
  3. Changed2 schema fields changedv2.3.0
    • changedInput schema / properties / image_mime_type / enum
      Previous value: -[
-  "image/jpeg",
-  "image/png",
-  "image/webp",
-  "image/gif",
-  "image/bmp"
-]New value: +[
+  "image/jpeg",
+  "image/png",
+  "image/webp",
+  "image/gif",
+  "image/bmp",
+  "image/tiff"
+]
    • changedInput schema / properties / images / items / properties / image_mime_type / enum
      Previous value: -[
-  "image/jpeg",
-  "image/png",
-  "image/webp",
-  "image/gif",
-  "image/bmp"
-]New value: +[
+  "image/jpeg",
+  "image/png",
+  "image/webp",
+  "image/gif",
+  "image/bmp",
+  "image/tiff"
+]
  4. First observedv1.3.0

TDQS

A4.1/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only provide openWorldHint, so the description must carry the behavioral burden. It discloses that an external vision model is used and that results come back as text, but it does not mention latency, failure modes, privacy implications of sending images externally, or output variability. This is meaningful but incomplete context.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is compact and front-loaded: purpose first, then usage restrictions, then input forms, then prompt advice. Every sentence contributes value, though the input-form list and prompt advice slightly overlap with schema descriptions. No wasted words overall.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For an 8-parameter tool with no output schema and no siblings, the description plus detailed schema covers target users, usage restrictions, input formats, prompt best practices, and output shape via output_format. Minor gaps remain around external-model reliability and potential sensitive-data considerations, but the core guidance is complete enough for correct invocation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100% and each parameter already has detailed semantic guidance (e.g., when to use image_ref, how output_format behaves). The description adds a useful summary of accepted input forms and prompt advice, but mostly restates what the schema already documents. Baseline 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description names a specific resource (image understanding) and a clear set of capabilities: OCR, scene description, chart/diagram interpretation, and image comparison. It also states the target audience (text-only models), which removes ambiguity about what the tool is for.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explicitly says when NOT to use the tool: multimodal models must not call it and should use their own vision capability instead. It also gives concrete prompt guidance (ask a focused question rather than 'describe everything'), which directly improves invocation quality.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools