A whole document, one MP3
Upload a PDF, a Word document (DOCX) or a PowerPoint deck (PPTX) up to 50 MB, or paste a long text: you get a single MP3 file (44.1 kHz, 128 kbps) to play or download.
Voice model · Deepnia
Coming soon to everyone
Deepnia Long-form — A whole document, read faithfully, in one MP3.
Deepnia Long-form is Deepnia’s own long-form narration tool: it turns a PDF, a Word document, a PowerPoint deck (up to 50 MB) or a long pasted text into a single MP3 file. The text is extracted in reading order, never summarized or rewritten; you review the full transcript and choose the voice before any audio is made. Deepnia Long-form is coming soon: the tool is already visible as a preview.
- PDF, DOCX, PPTX or text
- Up to 50 MB
- PDFs up to 500 pages
- One MP3
- Transcript to review
- Text read as written
Deepnia Long-form is coming soon: the tool already opens as a preview. The exact cost, recalculated from the approved transcript, will show before every creation.
- Document
- Transcript
- Voice
- MP3
At a glance
- Maker
- Deepnia
- Inputs
- PDF, DOCX or PPTX up to 50 MB, or pasted text
- Up to 500 pages
- Output
- One MP3 (44.1 kHz, 128 kbps)
- Review
- Full transcript, editable before the audio
- Voices
- English, French and Spanish narrators
What Deepnia Long-form does best
Faithful narration
The text is extracted in reading order, with no summary, addition or rewording. Only repeated headers and footers, page numbers, images and words split at line ends are removed; table rows are read cell by cell.
The transcript reviewed before the audio
You read and correct the full transcript, with search. Uncertain passages from text recognition are flagged with their page and a confidence level, and each one can be marked as checked.
A long job you can follow
The tool shows every stage: document analysis, text extraction, transcript to approve, audio creation, MP3 assembly. You can close the page: processing continues in the background and a notification tells you when the audio is ready.
Your choice of narrator
Pick a voice from a library of English, French and Spanish narrators, filter by gender and listen to samples first. The tool also shows a personal voices tab for a voice created with your permission.
Private files
Your upload stays a private file. 24 hours after the job ends, the source document and intermediate files are deleted; only the MP3 is kept.
How to use Deepnia Long-form on Deepnia
Upload the document
Drop a PDF, DOCX or PPTX (50 MB at most) or paste your text: the AI extracts the text in reading order.
Review the transcript
Fix the text, check the flagged passages, then save: every correction recalculates the cost.
Choose the voice and approve
Listen to samples, filter by language or gender, then approve the exact cost, which is locked at approval.
Get your MP3
The AI creates a single MP3 to download or open in your Library. You can also cancel the job along the way.
Deepnia Long-form on Deepnia: what you can set
- Inputs
- PDF, DOCX or PPTX in the document-to-audio workflow; PDF or pasted text in the voice tool with Deepnia Long-form selected
- Size
- Up to 50 MB per document and 500 pages per PDF
- Approved text
- Up to 1,000,000 bytes (UTF-8)
- Cleanup
- Repeated headers and footers, page numbers, images and split words removed; no summary or rewording
- Checks
- Search in the transcript; uncertain passages flagged with page and confidence level
- Voices
- English, French and Spanish narrator library, female / male filter, personal voices tab
- Output
- One MP3, 44.1 kHz, 128 kbps
- Progress
- Visible stages, background processing, notification, cancel option
- Privacy
- Private file; source and intermediate files deleted 24 h after the job ends
- Narration
- The same narrator from start to finish
Prompt ideas to try
Copy a prompt, open the tool and adapt it to your project.
Annual report for a nonprofit
- Female voice
- English
A 40-page annual report with a few tables of figures, for members to listen to before the general meeting.
Course notes to revise on the go
- DOCX
- Male voice
- English
30 pages of biology course notes to replay on the bus while revising for exams.
Training deck for a sales team
- PPTX
- Female voice
- English
A 25-slide training deck on welcoming customers, for new hires to listen to before their first week.
A chapter from a novel manuscript
- Pasted text
- Male voice
- English
The first chapter of a novel, pasted as plain text, to hear the pacing of the story before preparing an audiobook.
Newsletter for a hiking club
- Female voice
- Spanish
An 8-page monthly newsletter in Spanish, for club members who would rather listen than read.
Audio version of a user guide
- Male voice
- French
A user guide for a home appliance, in French, offered as audio to customers with low vision or who prefer to listen.
Tips for better results
- Use a PDF with selectable text where you can: scanned pages go through text recognition and produce more passages to check.
- Read the flagged passages first, then mark them as checked: that is where text recognition hesitated.
- Fix acronyms, figures and proper names in the transcript: the audio is made from the text you approve.
- Delete what should not be read, such as the table of contents, footnotes or bibliography, before you approve.
- Choose a voice in the document’s language: the library filters by language (English, French, Spanish) and by gender.
- Check how tables read in the transcript: each row is read cell by cell.
Good to know
- For a DOCX or PPTX, use the document-to-audio workflow; in the voice tool, upload a PDF or paste your text.
- Scanned pages go through text recognition: allow time to review the flagged passages before you approve.
- Your MP3 stays in your Library; the source document and intermediate files are deleted 24 hours after the job ends.
Deepnia Long-form or another model?
Deepnia’s other voice and music models, to pick the right one for your project.
| Model | Inputs | Settings | Best for |
|---|---|---|---|
| Deepnia Long-formThis model | Text, PDF, Authorized personal voice | — | A whole document, read faithfully, in one MP3. |
| Open TTS | Text, Authorized personal voice | Language, Speed, Expression tags | Text up to 5,000 characters with expression tags, speed and SRT captions. |
| MiniMax Speech 2.6 HD | Text, Authorized personal voice | Language, Speed | A voice-over up to 5,000 characters in your own cloned voice. |
| ElevenLabs | Text | Language, Stability, Expression tags | A short voice-over with expression tags, in 10 languages or auto. |
| Gemini Flash TTS | Text | Expression tags | A short voice-over with expression tags and automatic language detection. |
Frequently asked questions
Which files can I turn into audio?
PDFs, Word documents (DOCX) and PowerPoint decks (PPTX) up to 50 MB, in the document-to-audio workflow. In the voice tool, Deepnia Long-form takes a PDF or pasted text.
What is the maximum document size?
50 MB, and 500 pages for a PDF. The approved text can reach 1,000,000 bytes.
Is the text read faithfully?
Yes. The text is read as written, in reading order: only repeated headers and footers, page numbers, images and words split at line ends are removed, with no summary, addition or rewording.
Can I correct the text before the audio is made?
Yes. The full transcript is editable and searchable, and uncertain passages are flagged with their page and confidence level. Every correction recalculates the cost before approval.
What format is the audio?
A single MP3 file (44.1 kHz, 128 kbps), ready to download or open in your Library.
Can I close the page during the conversion?
Yes. Processing continues in the background and a notification appears when the audio is ready. You can also cancel the job.
Which languages can a document be read in?
The library offers English, French and Spanish narrators: pick a voice in the document’s language.
Is Deepnia Long-form available?
Deepnia Long-form is coming soon: the tool is already visible as a preview, marked “Bientôt disponible” (coming soon), and will open to everyone shortly. This page describes the workflow as the tool presents it.
Deepnia Long-form is coming soon
Deepnia Long-form is coming soon: the tool already opens as a preview. The exact cost, recalculated from the approved transcript, will show before every creation.