omni-vision-pro
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| GEMINI_MODEL | No | The Gemini model to use. Defaults to gemini-1.5-flash, but the server automatically falls back to GEMINI_FALLBACK_MODEL if unavailable. Example: gemini-2.5-flash. | gemini-1.5-flash |
| GEMINI_API_KEY | No | Your Google Gemini API key. If configured, Gemini is tried as the first image analysis provider. | |
| OPENAI_API_KEY | No | Your OpenAI API key. If configured, OpenAI is used as the second cloud provider after Gemini. | |
| OMNI_ALLOWED_ROOTS | No | Optional filesystem boundary for code and ZIP tools. Only paths inside these roots are allowed. Separate multiple roots with semicolons on Windows or colons on macOS/Linux. | |
| GEMINI_FALLBACK_MODEL | No | The fallback Gemini model to use when the primary model is unavailable. Defaults to gemini-2.5-flash. | gemini-2.5-flash |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| analyze_imagesA | Resolve one or more local/OpenCode image paths, resize them safely, analyze each in strict order with Gemini then OpenAI low-detail then local OCR, and return delimited JSON context. |
| read_code_contextA | Read a source file or directory without following symlinks, omitting vendor/build folders, binaries, and common secret files, then return a clean tree and contents. |
| read_zip_contextA | Inspect a .zip archive entirely in memory with path, size, ratio, secret, binary, and junk filtering; return a virtual file tree and text contents without extracting to disk. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
Each tool targets a completely different content type: images, source code/directories, and zip archives. There is no overlap or ambiguity between their purposes.
All tool names follow a consistent verb_noun pattern: analyze_images, read_code_context, read_zip_context. The naming is uniform and predictable.
With only 3 tools, the server is well-scoped and each tool has a distinct, focused responsibility. The count is at the lower end but well within the typical range for a focused utility server.
The server covers images, source code, and zip files, but lacks support for other common file types like PDFs, generic plain text, or non-zip archives. This leaves notable gaps in a general-purpose 'context extraction' server.