Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault

No arguments

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

CapabilityDetails
tools
{
  "listChanged": false
}
prompts
{
  "listChanged": false
}
resources
{
  "subscribe": false,
  "listChanged": false
}
experimental
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
inspect_pdf_linksA

Inspect Markdown/HTML references before PDF conversion, without writing files.

Reports local source paths, source existence, actual PDF candidates, nearby
PDFs and selection evidence. A unique same-stem PDF or recorded source can
be selected automatically; variants and competing sources need review.
For differently named legacy PDFs, inspect nearby files and project output
conventions; the tool cannot infer arbitrary correspondence from a name.
Supply pdf_targets mapping source filenames to actual PDF filenames, or
null to retain the source (e.g. README.md without a PDF). Relative mapping
paths resolve from file_path's directory. A target's .pdf-links.json can
persist a reviewed mapping, relative to that manifest's directory.
Non-Markdown/HTML targets such as .py files retain their original paths.
PDF chapter fragments are verified against the actual PDF. Ambiguous or
missing chapters require resolve_pdf_destination and a pdf_destinations
mapping from the original href to the chosen verified fragment.
resolve_pdf_destinationA

Locate a chapter in the actual target PDF without writing files.

Searches existing named destinations, exact normalized outline titles, then
complete text lines. target is the title, source anchor, or PDF fragment.
Returns status, physical 1-based page, printed page_label, evidence basis,
and candidates. Only a unique match returns a top-level fragment and uri.
For repeated titles, inspect the candidates and use page to narrow the
search or select a candidate's fragment. Never guess printed-page offsets.
For scans without usable text or bookmarks, use OCR to establish the page,
then validate page=N here. Text-line matches require context review.
Use the returned URI for PDF links, or pass pdf_destinations={original_href:
fragment} when converting Markdown/HTML. Recheck after the target changes.
markdown_to_pdfA

Convert a Markdown file (.md) to a styled PDF.

Uses markdown-it-py for parsing. Supports two rendering backends:
- engine="chromium" (default): Uses Playwright/Chromium for PDF output.
  Full MathJax SVG support — recommended for math-heavy documents.
- engine="weasyprint": Lightweight, good for simple documents without math.
Supports Chinese fonts, tables, code blocks, blockquotes, and page numbers.
Three color themes are available: white for printing, warm sepia for
low-glare reading, and a dark theme inspired by One Dark Pro Night Flat.
Preserves clickable web and local file links, including file:// URLs and
relative paths into other repositories, resolved from the source directory.
Call inspect_pdf_links first. Existing unambiguous derived PDFs replace
local Markdown/HTML targets; absent PDFs leave the source link unchanged.
Multiple PDF candidates require pdf_targets; other file formats stay intact.

Args:
    file_path:   Absolute path to the .md file.
    output_path: Absolute path for the output .pdf file.
                 If empty, the PDF is saved next to the source with the
                 same stem (e.g. /tmp/report.md → /tmp/report.pdf).
    engine:      Rendering backend — "chromium" (default) or "weasyprint".
    theme:       Color theme — "print" (white, default), "sepia" (warm),
                 or "one-dark-pro" (One Dark Pro Night Flat-inspired).
    pdf_targets: Reviewed source-to-PDF paths, or null values to keep sources.
                 Relative paths resolve from the source directory.
    pdf_destinations: Original hrefs mapped to fragments from resolve_pdf_destination.
    link_policy: "prefer-pdf" selects existing PDFs; "preserve" keeps sources.
html_to_pdfA

Convert an HTML file (.html) to PDF with print-aware styling.

Supports two rendering engines:

- ``chromium`` (default): Uses Playwright headless Chromium.  Pixel-identical
  to Chrome Print → Save as PDF.  Supports all modern CSS.
  Requires: ``pip install playwright && playwright install chromium``.
- ``weasyprint``: Lightweight, good for simple documents.
  Replaces emoji with font-styled spans.  May not match Chrome perfectly
  for display:flex / display:grid layouts.

Ordinary HTML keeps its original styles. Recognized portable analytics
reports receive a scoped A4 print profile for typography, cards, charts,
and source deduplication. Chromium also converts very wide tables to
labeled record cards. Three color themes are available: white for printing,
warm sepia for low-glare reading, and One Dark Pro for dark-room reading.
Preserves clickable web and local file links, including paths into other
repositories. Relative links respect the source directory or HTML base URL.
Call inspect_pdf_links first. Local Markdown/HTML links prefer verified
existing PDFs. Multiple editions require pdf_targets; references without
PDFs and other formats keep the source file. Only authored static anchors
are inspected; script-created links need an authored equivalent or review.

Args:
    file_path:   Absolute path to the .html file.
    output_path: Absolute path for the output .pdf file.
                 If empty, derived from the source stem.
    engine:      Rendering backend: ``"chromium"`` or ``"weasyprint"``.
    theme:       Color theme: ``"print"`` (white, default), ``"sepia"``
                 (warm), or ``"one-dark-pro"`` (dark).
    pdf_targets: Reviewed source-to-PDF paths, or null values to keep sources.
                 Relative paths resolve from the source directory.
    pdf_destinations: Original hrefs mapped to fragments from resolve_pdf_destination.
    link_policy: "prefer-pdf" selects existing PDFs; "preserve" keeps sources.
svg_to_pngA

Rasterize a self-contained SVG file to a PNG image.

The conversion runs locally on CPU. External file and network references
are blocked; embedded ``data:`` resources remain supported. Rendered output
is limited to 8192 pixels per side and 32 million total pixels, then
validated and atomically published so a failure cannot replace an existing
destination with partial bytes.

Args:
    file_path:       Absolute path to the .svg file.
    output_path:     Absolute path for the output .png file. If empty, the
                     PNG is saved beside the source with the same stem.
    scale:           Positive scale factor up to 16. Cannot be combined
                     with output_width or output_height.
    output_width:    Optional positive pixel width. If output_height is
                     omitted, the aspect ratio is preserved.
    output_height:   Optional positive pixel height. If output_width is
                     omitted, the aspect ratio is preserved.
    background_color: Optional CSS color for otherwise transparent pixels.
pdf_to_textA

Extract plain text from a born-digital PDF using PyMuPDF.

**born-digital PDF only** (text that can be selected/copied with a mouse).
Scanned-image PDFs will return an empty string — use ``ocr_document``
for those.

Args:
    file_path: Absolute path to the .pdf file.
    save_text: If True (default), also writes a .txt file alongside the PDF.

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A4.3/5.0

Scored across 6 tools

Disambiguation4/5

The conversion tools are clearly separated by input format: markdown_to_pdf, html_to_pdf, pdf_to_text, and svg_to_png each target a distinct source type. resolve_pdf_destination and inspect_pdf_links both deal with PDF linking, but their descriptions distinguish anchor resolution from pre-conversion link inspection.

Naming Consistency4/5

All tools use snake_case and the four conversion tools follow a predictable source_to_target pattern. The two helper tools use verb_noun naming, which is readable and appropriate, though it means the set is not a single uniform verb_noun convention throughout.

Tool Count5/5

Six tools is well-scoped for a format conversion server, with each tool covering a distinct conversion or PDF-link support operation. There is no obvious redundancy or bloat.

Completeness3/5

Core paths such as Markdown/HTML to PDF, PDF to text, SVG to PNG, and PDF link resolution are covered. However, pdf_to_text explicitly directs agents to ocr_document for scanned PDFs, but no such tool exists in this set, creating a dead end; common reverse conversions such as PDF to Markdown/HTML are also absent.

Maintenance

ActivityNo data
ResponsivenessOutdated data