markitdown-ocr-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| OMLX_URL | No | The URL of the oMLX server (OpenAI-compatible endpoint). | http://127.0.0.1:8080/v1 |
| OMLX_API_KEY | No | API key for the oMLX server. Auto-read from ~/.omlx/settings.json if not set. | |
| OMLX_OCR_MODEL | No | OCR model to use. Auto-discovered from /v1/models if not set. | |
| OMLX_OCR_PROMPT | No | Prompt used for OCR extraction. | OCR: |
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": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| inspect_pdfA | Classify each page of a PDF as text / scanned / mixed / blank. Uses the PDF text layer and image geometry only — no OCR is run. Use this first to decide whether (and which pages need) OCR. |
| ocr_pdfA | Convert a PDF to Markdown, OCR-ing scanned/image pages via oMLX. Text-layer pages go through MarkItDown directly; scanned pages are
rendered and OCR'd by the vision model. If |
| omlx_modelsA | Check oMLX health: available models and the resolved OCR model. |
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 has a clearly distinct role: inspect_pdf classifies pages, ocr_pdf performs conversion, and omlx_models checks OCR service health. There is no functional overlap or ambiguity between them.
All names use lowercase snake_case and follow a pattern of verb_noun (inspect_pdf, ocr_pdf). omlx_models breaks the verb_noun pattern but is still readable and consistent in style, making it a minor deviation.
With only three tools, the server is tightly scoped to the PDF-to-Markdown OCR workflow. Each tool serves a necessary purpose without redundancy or bloat.
The server supports the full workflow: inspecting PDFs to identify OCR needs, converting with optional page selection, and verifying OCR model availability. No critical missing operations are apparent for its stated purpose.