For free chapter MP3s or a single M4B from a PDF with selectable text, start with Audiobook.tools. If you need to read and correct a scan before making chapter files, Lacuna documents both steps, with downloads on paid plans. If you already have selectable English text and need to fix the reading order or set your own section breaks, start in our PDF workspace.
Start with the result you need
Decide how you want to listen. Offline playback inside an app may keep the audio in that app. Separate chapter MP3s let you move individual tracks; a chaptered M4B keeps the whole book in one file with navigation in compatible players. A ZIP is a bundle of files. If you need the full text, choose a read-through rather than a podcast or summary.
Scroll the table to see all columns.
| Product | Where it runs | Text and chapters | Scanned pages | Downloads | Free or paid? |
|---|---|---|---|---|---|
PDF to Audiobook (this site) | PDF in your browser; selected text sent for narration. | Edit text; rename, split, merge or skip sections. | Run OCR first. | Chapter MP3s or ZIP; reviewed TXT. No M4B. | Free text review. Trial when available; paid character plans. |
| PDF2Audio.ai | Upload for extraction and narration. | Automatic chapters. Text edits and manual cuts unconfirmed. | OCR in PDF to MP3. Book mode unconfirmed. | Whole-book or chapter MP3s. | Signup allowance; pay per PDF or buy credits. |
| FreeTTS.org | Upload for extraction and narration. | Rename or skip chapters. Text edits unconfirmed. | OCR on Pro and Creator. | Chapter MP3s or ZIP. Creator lists M4B. | Tagged audio on Free. Our ZIP download prompted for an account. |
| Audiobook.tools | Extraction and narration both in your browser. | Preview cleaned text; rename or skip. Direct edits unconfirmed. | Run OCR first. | Chapter MP3s or chaptered M4B. | Free exports; no account. |
| Lacuna | Upload for extraction and narration. | Edit or skip chapters. Manual splits and merges unconfirmed. | Reads image-only pages. | Chapter MP3s, WAVs or ZIP. | Free listening; downloads need a subscription. |
Compare formats, limits and total prices
Compare the same job: 100,000 cleaned characters, one narration pass, personal listening, and a file you can keep. Count any charges for reading the PDF, making corrections that need new audio, and downloading the file. Check the smallest plan or purchase that covers the job. Compare the full annual payment with the price of one month, and check the final amount for local currency and taxes.
At PDF2Audio.ai's one-off rate of US$0.45 per 10,000 characters, that job is US$4.50 at list price. On Lacuna, it uses 5,000 speech credits plus 5 credits per PDF page, and file downloads still require a subscription. In Studio, a model priced at one credit per character would use 100,000 credits before revisions. Check the model rate and shared allowance on ElevenLabs pricing. These are calculations from published rates. The free workflow guide explains how to distinguish a free preview from a free export.
Limits that can change your choice
- PDF to Audiobook: Review text and export TXT without signing in. The editor accepts PDFs up to 20 MiB and 600,000 extracted characters. The pricing page lists the signup trial allowance for short excerpts. It also lists US$9 for 25,000 characters valid for 180 days and US$29/month for 100,000 with no rollover, along with current purchase availability. The privacy policy explains what is sent when you confirm narration.
- PDF2Audio.ai: PDFs are limited to 10 MB and about 2 million characters. Its MP3 tool documents OCR for up to 30 scanned pages. Book mode inserts chapter markers automatically; the guide describes roughly ten-minute sections on longer files. Its published comparison lists up to 30,000 signup characters. The Book guide advertises about 20 minutes of preview for a longer first file; downloading that preview is unconfirmed. One-off conversions start at US$2.99.
- FreeTTS.org: The PDF tool lists one free PDF per day (up to five pages), a 5,000-character limit per chapter and an audio tag. It uses bookmarks or divides the document into equal parts. PDFs have a 50 MB cap. Pro lists up to 500 pages per PDF at US$19/month; Creator lists up to 2,000 pages at US$39/month. Creator pricing lists M4B; the PDF FAQ describes it as planned for Pro.
- Audiobook.tools: The PDF tool previews cleaned text and uses bookmarks or headings. The separate Script Cleaner handles copied text; direct edits and custom chapter breaks in the PDF tool are unconfirmed. First setup downloads about 93 MB in CPU mode or 327 MB with WebGPU, and long books need enough device memory.
- Lacuna: The PDF tool allows up to 500 pages and 100 MB. Free accounts receive 120 credits a month. Reading costs 5 credits per page before review; narration costs 5 per 100 characters. Chapters longer than about 30 minutes are split. Lacuna says it deletes the PDF after reading it and keeps extracted text until you delete it. Credit packs alone do not unlock downloads; check the subscription options.
Where each option fits
- One long book without a subscription: PDF2Audio.ai Book mode offers a one-time purchase and whole-book or chapter MP3s. Compare its automatic chapter boundaries with your table of contents. For a short PDF, FreeTTS.org generated both chapters in our anonymous test, but asked for a free account when we tried to download the ZIP.
- Keep the PDF and the script on your device: Audiobook.tools documents local narration as well as extraction. If you need a different voice or more control over the setup, see the local workflow guide.
- Fix a difficult layout before generating: Try our two-column PDF example. It includes the PDF, raw and corrected scripts, and matched before-and-after audio. If you cannot select a word on the page, start with the scanned PDF guide or consider Lacuna for scan recognition and editing in one workflow.
When a reader app or production editor fits better
- ElevenReader suits in-app listening. Its PDF guide documents scan recognition and says imported PDFs cannot be exported as standalone MP3s.
- ElevenCreative Studio suits a manuscript that needs paragraph-level edits and chapter organization. Its documentation covers PDF import and MP3 or WAV export by chapter, ZIP or whole project. PDF import alone does not establish OCR.
- NaturalReader Personal offers OCR in batches of up to 50 pages and paid MP3 downloads in batches of up to 20 pages or 50,000 characters of pasted text. Free Voices cannot be exported as MP3. Check the Personal plan terms if you plan to share or publish audio.
Test the same three PDF cases
Use the test files ZIP or open the files below. These short English documents were made for this comparison. The set has three cases and four PDFs because the scan comes both before and after OCR.
- Two ordinary chapters: open the bookmarked PDF and keep the reference text beside it. Record the section labels and boundaries. The two chapters also span two pages, so the count alone does not prove that the tool used the bookmarks.
- Two columns and a layout note: open the original PDF. Check whether the text jumps between columns and reads the running header or note. The worked example shows the raw text, the manual correction and two clips with the same voice.
- A scan before and after OCR: open the image-only PDF first. Then use the same page with a text layer. Compare the OCR text with the reference text before making audio.
Start a fresh document for each file. Save the raw extract before changing it and record the section count, any missing or reordered text, and how you corrected it. For audio, record the files you actually downloaded and any account, credit or plan limit you hit. An in-app play button does not establish that a file can be exported.
What we observed in our workspace
On November 8, 2026, we imported the sample PDFs in an isolated local preview without an account or a cloud narration request.
- Two ordinary chapters: the editor created two page sections, labeled Page 1 and Page 2. The text kept both chapter headings and the values 12 minutes, 6 crates, 15 seconds and 4 knots. The workspace TXT export shows the page labels and both sections.
- Two columns: the editor reproduced the interleaving in the raw extract. We restored the reading order by hand, split the corrected first page into two sections, and exported them as TXT. The clips in the worked example use the same offline voice and rate.
- Scan: the image-only PDF showed “No selectable text was found.” The same page after OCR imported as one section. The workspace TXT export matched the reference after ignoring line breaks and extra spaces, including 18.5 meters and 42 seconds.
Evidence and testing scope
We checked the linked official product pages on November 8, 2026. PDF to Audiobook publishes this comparison and is one of the products listed. See who operates the site.
The workspace observations above come from the same downloadable files. The scan's OCR preparation ran separately before import. We did not submit a cloud narration job in our workspace or test checkout or deletion.
The feature table cites official product information. We tested the same two-chapter PDF in both tools below, in separate Chromium sessions without signing in or paying. We also opened the two-column file and both scan versions in Audiobook.tools.
Audiobook.tools: text preview and chapter MP3s
We used the PDF tool in its browser-based mode. Requests were limited to the tool and its font and model assets; the test did not upload the PDF.
- Ordinary chapters: the tool reported both bookmarks and gave both sections the matching chapter titles. The previews kept the chapter bodies and turned the numbers into words. The downloaded ZIP contained two MP3s, which decoded successfully at about 14 and 16 seconds. We did not test M4B export.
- Two columns: the preview omitted the sample's print layout note but kept the running header and still mixed the left and right columns. It jumped from “a brass lamp” to “Once the lamp was lit.” This is a result for the sample above, not a rule for all two-column PDFs.
- Scan before and after OCR: the image-only file prompted us to run OCR. The searchable version produced one section with the correct heading, and its preview preserved the values 18.5 meters and 42 seconds. For the two-column and scan files, we checked text only.
FreeTTS.org: playback and a ZIP sign-up prompt
We uploaded only the two-page chaptered sample to the PDF tool. It kept both chapter titles, and both chapters played in the browser. The read-along text retained 12 minutes, 6 crates, 15 seconds and 4 knots.
When we chose Download all as ZIP, the tool asked us to create a free account. We stopped there and received no ZIP or MP3. We did not try the individual chapter download buttons or run the two-column or scan files through this tool.
Tools not tested with these files
PDF2Audio.ai and Lacuna remain documentation-only in this comparison. We did not upload the samples to them or check their chaptering, OCR or downloads.
Before committing to a book, try a page you have permission to use that contains the hardest layout or pronunciation in your document. Check the text order and chapter boundaries, confirm what the file download includes, and compare the quoted cost using the same input and output.