AI Glot translates files, and a Google Doc is not a file until you export it. That is the only extra step. Export to Word or PowerPoint, translate, upload the result back to Drive.
This guide shows the loop for a Doc and for a Slides deck, and when Google’s own translator is enough.
Which option should you use?
| Good for | Where it stops | |
|---|---|---|
| Google Docs, Tools, Translate document | A short Doc, once. It creates a translated copy as a new document | One document at a time, and you pick a language and nothing else: no glossary, no scope, no plan to read first |
| DeepL | One file, quickly. It takes Word and PowerPoint files up to 30 MB | You export first, as with any file route, and then it works one file at a time with a glossary that swaps terms without understanding why they are fixed |
| ChatGPT, Claude or another assistant | A few paragraphs | You paste text out and back in, and carry your glossary into every chat. A long Doc uses up your usage limit |
| AI Glot | Long Docs and decks, several languages, a glossary that must hold, documents you update often | The export and the upload add two small steps |
| An agency or a freelancer | A document that goes to customers or gets signed | Slow, and priced per word |
Google’s translator is fine for a one-page Doc. The file route pays off the moment the document is long, the same words must appear everywhere, or you need more than one language.
The loop
- 1ExportFile, Download, .docx or .pptxIn Google
- 2TranslateUpload, one sentence, a glossaryIn AI Glot
- 3Read the planParts, languages, costFree
- 4ApproveThe only step that spendsSpends credits
- 5Back to DriveUpload the file, open with Docs or SlidesIn Google
Nothing is spent until you approve. Everything before step four is free.
Before you export
Do three small things, and the result is cleaner.
Accept or reject suggestions. Text under a tracked change is left alone by the translation, because translating a revision would mean deciding for the author whether to accept it.
Resolve the comments you no longer need. Comment threads are left as written.
Look for text inside images. A diagram pasted as a picture has no text for the engine to read. Keep a list of those labels to redo by hand.
How to do it, step by step
1. Export the file. In Docs, choose File, Download, Microsoft Word (.docx). In Slides, choose File, Download, Microsoft PowerPoint (.pptx).
2. Upload it to AI Glot in the app. You can try a small file first, with no account, in the Word translator or the PowerPoint translator.
3. Say what you need in one sentence.
Translate this document into Spanish.
Leave the contact table on the last page in English.
The sentence decides what gets translated. Rules about wording, such as “use the informal tone”, go into the second box at approval, because they are applied to each paragraph as it is written.
4. Add a glossary for names, products and the words you always use the same way. How to build one.
5. Read the plan and choose a quality level.
Lite gives about three times more words per credit and is faster: right for an internal Doc. Standard is for what a customer reads. One credit is one word in Standard.
6. Approve and download.
7. Back into Google. Upload the translated file to Drive, then open it with Google Docs or Google Slides. It becomes a normal Google file again, ready to share.
One Doc, several languages
The file names are an example. Each language is its own file, so each one becomes its own Google Doc.
Three ways to run it
In the app is the path above.
With your AI agent: connect AI Glot as an MCP server. The agent uploads the exported file, writes the plan, updates your glossary and downloads the result.
With the API or the command line: if you export Docs automatically, for example every policy page that changes, you can send each file straight to translation. See the API documentation.
Check the result
Open the translated Doc in Google. Check a heading, a table and the first page of a long section. Keep the original until someone who speaks the language has read the parts that matter, and compare the two side by side once before you share the translated version.
For spreadsheets, the same loop works with an Excel export: see how to translate an Excel file. Pricing lists what a larger volume costs.
The export steps above are Google’s menu path at the time of writing and can change. DeepL’s limit comes from its own documentation. File names in the diagrams are examples. Your plan shows the real figures for your own file, and it is free to read.
