mdfi-connect
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MDFI_API_KEY | Yes | Required. Create one on the Connect page of your mdfi dashboard. | |
| MDFI_API_URL | No | Optional. The base URL of the mdfi cloud API. Defaults to https://api.mdfi.ai. | https://api.mdfi.ai |
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
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| read_documentA | Read a local PDF, Word, PowerPoint, Excel, HTML, email or image file as clean text. Handles scans and complex tables, and marks anything uncertain. Prefer this over reading such files directly. |
| read_pagesA | Read specific pages of a document already opened with read_document. Pass pages (e.g. "3,7,9-11") or section (a heading from the outline), not both. Free: never charges. |
| export_tableA | Write one table from a document already opened with read_document to a CSV file in the working directory (default ./mdfi-tables/-.csv, UTF-8 with a header row). The table id comes from the table's label in the page text, e.g. [table p13-t1 — …]. A table with notes gets a .notes.txt file beside the CSV; never overwrites unless overwrite=true. Only once the document is ready. Free: never charges. |
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
read_document and read_pages both read document content, but the descriptions clearly distinguish opening a document from reading specific pages of an already-opened document. export_table is clearly separate, targeting CSV extraction of a specific table.
All three tools follow a consistent verb_noun snake_case pattern: read_document, read_pages, export_table. The naming is predictable and readable throughout.
Three tools are well-scoped for a document-reading and table-export server, with each tool covering a distinct stage of the workflow. There is no obvious filler or redundancy.
The surface covers opening documents, reading targeted pages/sections, and exporting tables, which is coherent for the apparent read-only document-access domain. Minor gaps might include document metadata or searching across documents, but the core workflow is complete.