You have a .tex file. A co-author who does not use LaTeX needs to comment on it, a course wants the submission as a Google Docs™ document, or a group is drafting in Docs and the maths lives in your file. Docs cannot open .tex, so the question is how much of the file survives the move.
This post walks the path that exists today with the free LaTeXFlow add-on, and then says plainly where that path stops. Nothing here is a promise about a feature that does not exist yet.
1. Two different problems in one file
A .tex file holds two kinds of content, and they move differently.
- Prose. Paragraph text is plain text. Copy and paste moves it, and Docs keeps it as text.
- Maths. Everything between LaTeX math delimiters —
$x^2$,\[ … \], and the rest — is also plain text after a paste, but it looks like LaTeX rather than like an equation. This is the part LaTeXFlow handles: it finds those spans and turns each one into a rendered equation image.
Everything else in the file — sectioning commands, lists, tables, citations, figures — is a third kind of content, and it is where the path ends. Section 4 is about that.
2. The free path, step by step
Step 1 — copy the body of the file. Open the .tex in any text editor. Select from the line after \begin{document} to the line before \end{document} and copy. The preamble — \documentclass, \usepackage lines, macro definitions — has no meaning in Docs; leave it behind.
Step 2 — paste into a Google Docs™ document. Paste as plain text (Ctrl+Shift+V on Windows, Cmd+Shift+V on a Mac) so no stray formatting comes along. You now have a document full of readable text with LaTeX commands sitting in it as characters.
Step 3 — open LaTeXFlow. In the document: Extensions → LatexFlow → Open Equation Panel. If this is the first run, the add-on shows a two-page consent screen about optional data collection; answer it either way, every feature works the same. The install walkthrough covers that screen in detail.
Step 4 — Scan. Switch to the Scan tab and press 🔍 Scan Document. The sidebar lists every span it recognised as an equation, one card each, with a preview. Look through the cards. A card that is not maths — a price with a stray dollar sign is the usual case — can be skipped.
Step 5 — Convert All. Press ⚡ Convert All. Each detected span is replaced in place by an equation image. The LaTeX source is kept in the image’s alt text, so ↩ Revert to LaTeX on the same tab turns any image back into editable text later.
Two things to be clear about. The result is an image per equation, not a native Docs equation — the objects Docs makes with Insert → Symbols → Equation are a different thing and this path does not produce them. And the prose around the equations is exactly what you pasted: plain text, in one font, with the LaTeX commands still visible wherever they were. Steps 3 to 5 changed the maths and nothing else.
3. What Scan detects, and what it does not
This is the part worth reading before you paste a long file, because it decides how much editing you do first.
The scanner detects delimited maths only. It looks for text wrapped in one of four delimiter pairs — $…$ and \(…\) for inline, $$…$$ and \[…\] for display — plus, when you tick Also detect [ … ] in the panel, the bare square brackets that ChatGPT produces when it strips the backslashes from \[ … \]. That option is off by default in the add-on. Maths that is not inside one of those pairs is not an equation to the scanner. This is a deliberate boundary rather than a gap waiting to be filled: a detector loose enough to catch bare \frac{1}{2} also catches file paths, code, and any sentence that happens to contain a backslash, and it would start flagging documents that were fine.
In a .tex file, three consequences follow.
Environments are not delimiters. An equation written as
1
2
3
\begin{equation}
E = mc^2 \label{eq:energy}
\end{equation}
is not detected, because \begin{equation} is not one of the four pairs. The same holds for align, align*, gather, and multline. To bring it across, wrap the body in display delimiters and drop the label — Docs has no equation numbering for a label to point at:
1
\[ E = mc^2 \]
Aligned bodies need aligned. When the body uses & alignment points and \\ line breaks, put it in an aligned block inside the display delimiters. This is the same shape the add-on’s own symbol palette inserts, so it is the shape the renderer expects:
1
2
3
4
\begin{align*}
f(x) &= x^2 + 1 \\
g(x) &= 2x
\end{align*}
becomes
1
2
3
4
5
6
\[
\begin{aligned}
f(x) &= x^2 + 1 \\
g(x) &= 2x
\end{aligned}
\]
One \[ … \] block renders as one image with both lines in it.
Inline maths cannot span a line break. Many .tex files are hard-wrapped at 80 characters. If an inline $ … $ pair happens to straddle a wrapped line, the two dollar signs sit on different lines and the pair is not detected — the inline pattern stops at a line end on purpose. Join the line in the text editor before you copy. The three other delimiter pairs are not affected.
If you find a paragraph after the scan that should have been an equation, you do not need to go back to the file: put the cursor on it, press Tag as $…$ on the Scan tab, and scan again. With a selection it wraps the selection; with only a cursor it wraps the whole paragraph.
4. What does not carry over
Everything in the file that is structure rather than text or maths arrives as plain text with its command still attached. \section{Method} is the six characters of a heading and a command name; \begin{itemize} and each \item are words in a paragraph; a tabular block is rows of & and \\; \cite{knuth84} and the bibliography are a key and a list; \label and \ref are a name with nothing to resolve it; \footnote{…} is the note inline in the sentence; \includegraphics{fig1.pdf} is a file name, since the image itself was never in the .tex. For each of these you apply the Docs equivalent by hand — heading styles, a bulleted or numbered list, a Docs table, a footnote from the Insert menu, the image file from disk — and delete the command text. On a short document this is a few minutes. On a thesis chapter it is most of an afternoon, and it is the honest description of where this path stops.
5. One line to close
We are looking at a proper .tex import that keeps the structure — headings, lists, tables, footnotes — so that step 4 is not done by hand; if that matters to you, say so on the support page and it helps us decide what to build next.