Practical Guide to Getting the Most Out of InsertMyComments.com

Inserting comments from a PDF into its source document will never be a 100% perfect operation. The very act of copying text from a PDF, even with professional tools like Adobe Acrobat Pro, already presents difficulties: content gets jumbled, formatting is lost, or boxes, numbering, or footnotes are omitted. And that’s just when trying to extract plain text. Imagine the challenge of precisely inserting that content back into another document.

On the other hand, not all source documents actually contain the visible text from the PDF. Some structures generate content “on the fly”: calculated fields in Word, cross-references in XML or LaTeX, text in attributes, metadata, or macros… All these cases can prevent the original fragment from being found or faithfully reconstructed.

InsertMyComments is no exception to these structural limitations. Like any tool that interacts with PDFs and complex documents, it can face challenges both when extracting highlighted text correctly and when inserting it into its destination. What matters is that each of these cases is analyzed and improved over time.

To follow that progress, you can check out a scope infographic, where you can get a sense of the current status of each PDF and source document combination through color, position, and key tags.

This very article is updated in parallel: each improvement to the system is reflected both in the infographic and in the content of this guide, so you always have a reliable and up-to-date reference. Below you'll find two complementary sections:

  1. A clear explanation of the two workflows offered by InsertMyComments:
    — one more manual, where you control every detail,
    — and another AI-assisted, which automates much of the process.
  2. A practical walkthrough, case by case, depending on the type of source document you’re using (Word, XML, LaTeX, etc.), with concrete tips to maximize accuracy in each scenario.

1. Who does the editorial work: you or the AI?

InsertMyComments always automates the insertion; the difference lies in who polishes the comments beforehand.

Path A – Manual preparation, automatic insertion (no AI)

This mode is designed for those who act as reviewers or editors: people who didn’t write the comment but refine it before insertion. Their task is to clean up the comment content and correctly prepare the highlighted text, so the tool can apply the changes precisely, without any AI intervention.

Here, InsertMyComments acts as an automatic paster: it copies and replaces exactly what you indicated, without reinterpreting anything.

✅ Advantages

  • Literal and predictable insertion.
  • Maximum processing speed.

🧩 Considerations

  • Requires manual preparation of each comment and each highlight.
  • No ambiguity allowed: the text to be replaced must be exact and unique.
  • No automatic interpretation: the system doesn’t rephrase or correct.
  • The result quality depends entirely on how well the case is prepared.

✍️ How to properly prepare comments

Comment box

It should contain only the exact word or phrase that must replace the highlighted text.

Examples:

  • Maybe better: girl?
  • girl

Highlighted text

You don’t need to extend it just because you're afraid of too many matches: the system automatically expands the context to find a unique instance — even without AI.

But if you expect the change to affect the grammatical surroundings (such as verb agreement or determiner choice), then expand both the highlight and the comment yourself.

Examples:

  • “team” → “players”
  • “the team is” → “the players are”

Path B – AI-assisted preparation, automatic insertion

This mode is intended for those who prefer to skip the manual preparation of comments and let the system act as a pre-editor. InsertMyComments with AI interprets what the author intended, cleans the comment, and automatically corrects grammatical agreement errors (gender, number, articles, prepositions…) in the context of the insertion.

However: even if you don't need to prepare anything beforehand, you must review the result afterwards.

✅ Advantages

  • Time-saving: you don’t need to edit comments or adjust highlights.
  • The system automatically corrects grammatical aspects in the insertion context.
  • Ideal for processing many comments at once with lighter post-review.

📎 Considerations

  • Processing may take a bit longer, especially with a high number of comments.

✍️ How to properly review the comments

In both the manual and AI-assisted modes, a detailed technical report is generated, where each comment is marked with a letter according to the type of intervention applied by the AI:

  • A: literal insertion with no changes.
  • B: the highlighted context was expanded.
  • C: the comment content was filtered.
  • D: grammatical adjustments were applied.

This report is also generated in manual mode, but it’s especially useful here because it helps you prioritize the review:

👉 You can focus on comments with high-level intervention (B, C, or D) and skip those that were inserted literally and directly (type A).

2. What you can do today depending on your document type

InsertMyComments is based on a simple principle: it relies on a PDF with highlighted comments and automatically applies the changes to an editable source document that you upload.

This section explains, step by step, which formats are accepted in each role (comment carrier and editable source) and what conditions must be met for everything to work precisely.

2.1. Comment carrier document: PDF only (for now)

Currently, InsertMyComments only accepts PDF as the base format for comments. It is the standard medium in professional editorial environments, especially in proofreading, galley corrections, or external review workflows.

Other formats like Word, Google Docs, LibreOffice, or similar also allow inserting comments, but our tool does not yet support their automatic processing. For example, InsertMyComments still cannot read comments directly from a .docx file and apply them to that same document.

We focus on highlighted comments in PDF because they are the most common in intensive annotation processes (dozens or hundreds of edits), whereas native comments in Word or other editors are typically used for occasional author-editor reviews.

2.2. Type of comments: highlights only

InsertMyComments works exclusively with comments based on highlighted text.

These are the most stable, widely compatible, and visually clear.

Other types of annotations (text boxes, callouts, strikethroughs, underlines, arrows, freehand drawings, stamps…) cannot be processed yet.

If you already have comments in another format, you can convert them manually or using external tools such as:

  • pdfannots2highlight.py (Python)
  • Export functions from Skim, Xodo, or Adobe Acrobat Pro

We are working to gradually expand compatibility.

2.3. Editable source document

The changes are applied to a source document that you upload. We currently accept:

  • Word (.docx): fully supported and robust. This is the most stable and accurate format, compatible with nearly all editorial workflows.
  • Structured XML: we support widely used schemas such as DocBook, TEI, JATS, and even well-formed HTML. The engine detects tags and preserves structure.
  • 🔴 InDesign / IDML: for now, InsertMyComments cannot operate directly on .indd files, but you can export your project from InDesign as IDML (which is XML) and apply the changes there.
    • ➤ This is a one-way flow: once processed, you’ll need to reimport the IDML into InDesign manually.
    • ➤ We’re exploring ways to automate this, but it’s not possible yet.
  • 🟡 LaTeX (.tex): you can upload a .tex file directly, and it will work as long as the highlighted text is in clean areas, without commands or macros.
    • ➤ If the text appears inside \textbf{} or other instructions, the match may fail.
    • ➤ As an advanced alternative, you can convert the .tex to XML (e.g., using LaTeXML) to make insertion easier.
  • 🔵 TXT (.txt): plain text is also supported as a source document. Although it’s rare for a .txt file to originate a commented PDF and need retroactive editing, it works like any other format.
    • ➤ The text must exactly match the highlighted portion from the PDF.
    • ➤ There are no interfering tags, but no helpers either: any space, accent, or line break can cause the insertion to fail.

📎 General tip: make sure to upload a continuous and clean document. If your source is split into multiple files (as in DITA projects or fragmented XML), merge them beforehand into a single file before using the tool.

This can be done easily from the Windows command line (CMD):

copy /b *.xml merged_file.xml

This will merge all the .xml files in the folder (in alphabetical order) into a single file called merged_file.xml. It also works with .txt, .tex, or other plain text formats.

2.4. Alternative editors and special formats

In addition to Word or XML, you can also work with other widely used editors, as long as you know how to integrate them into the workflow. All of the following support exporting to .docx and/or .pdf, so you can follow the same flow as in the standard case:

  • Google Docs, LibreOffice, OnlyOffice, Zoho Writer…
    • ➤ Download your document as a .docx, create a .pdf with comments, and apply the changes to the original .docx using InsertMyComments.
    • ➤ You can then re-upload the corrected document to your original platform if needed.
  • 🔴 Markdown (.md):
    • ➤ There is no direct support for .md, but you can easily convert it to .html or paste it into a clean .docx file to use as a source.
    • ➤ Useful in technical contexts or quick documentation workflows.

3. Conclusion and next steps

This guide will continue to evolve over time, just like the system itself: step by step, comment by comment.

If you work in publishing, technical, or academic fields and deal with documents and annotations, we hope InsertMyComments becomes an invisible yet effective ally.

And if something doesn’t work as expected, let us know: many of the improvements you see here started with a simple question from a user like you.

← Back to the main page