Merge files into one Markdown

Adding and ordering files

Three ways in, all local:

• Add files, pick one or many text files.
• Add folder, pick a folder and every supported file inside is queued (subfolders included).
• Drag & drop, drop files or a folder onto the page.

Each file becomes a row. Drag a row, or use the up and down arrows, to set the order; the merged document follows that order top to bottom. Remove a row with the × button.

Supported formats input

This is a text-merge tool. It accepts:

• .md / .markdown / .txt, passed through as-is.
• .csv, converted to a Markdown table.
• .srt / .vtt, converted to a clean transcript.

The CSV and subtitle conversions use the same engines as the dedicated CSV to Markdown and SRT/VTT to Markdown tools. PDF, DOCX, Excel, and notebook files are not part of this tool; convert those first with their own converter, then merge the Markdown.

Headings, separators, and output output

By default each file is introduced with its name as an ## heading and files are separated by a horizontal rule. You can switch headings off, use a plain blank line as the separator, or include the file's full folder path in the heading (handy after a folder drop).

The merged Markdown is shown in a live preview, rendered through iLoveMD's sanitized engine so it is safe even for files you did not write. Use Copy to put it on your clipboard, or Download .md to save it as merged.md. A running word, character, and GPT-4o token count helps you see how much context window the result will use.

A worked example: three files in, one document out

Say you are preparing a briefing from three sources: notes.md (meeting notes), metrics.csv (a small spreadsheet export), and interview.srt (subtitles from a recorded call). Drop all three in, drag them into that order, and the merged document comes out as:

## notes.md

(your notes, unchanged)

---

## metrics.csv

| Month | Signups | Churn |
| ----- | ------- | ----- |
| May   | 214     | 3     |

---

## interview.srt

(a clean transcript, timestamps removed)

The CSV became a real Markdown table and the subtitles became readable prose, so the single file reads like one coherent document instead of three pasted fragments.

Heading levels when files collide

The classic merge problem: each source file often opens with its own # top-level heading, so a naive concatenation produces a document with five competing titles. The default settings here are chosen to sidestep that:

• Filenames become ## section headings, one level below a top-level title, so the merged document has a clean hierarchy even when the sources disagree.
• Horizontal-rule separators mark where one file ends and the next begins, which keeps skimming honest.
• Headings inside your files are passed through as written. The tool does not renumber them, so if a source file uses # internally, consider whether you want the filename headings on (they give the document its structure) or off (when each file already titles itself).

Assembling context for an AI prompt

The most common use of this tool is bundling reference material for a language-model conversation: several docs pages, a transcript, a data table, one paste. Three habits make the bundle work harder:

• Order deliberately. Put the material the model should weight most either first or last; the middle of a long context is where detail is most often overlooked.
• Keep the per-file headings on. When the model can see ## interview.srt, it can attribute what it quotes, and you can ask questions like "what did the interview say about pricing?"
• Watch the token count. The running GPT-4o count below the preview is the same exact tokenizer as our Token Counter; if the bundle is oversized, trim files rather than letting the model silently lose the middle.

Transcript sources merge in cleanly: subtitle files convert on the way in, and the YouTube transcript tool produces Markdown that drops straight into the queue.