Tools

BookForge converts your ebooks and PDFs into natural-sounding audiobooks on your own computer — everything runs locally by default, and cloud AI is optional, never required.

1. Install & first run

1. Run the installer (see the virus-warning note above — it is a false positive). It is a single large file because the app bundles its own Python runtime.
2. On first launch BookForge unpacks and sets up its audiobook engine. This takes a few minutes the very first time — a progress bar at the bottom shows how it is going, and you can keep using the app while it finishes.
3. A short guided setup then walks you through optional extras: AI (a bundled local model, Ollama, or a Claude/OpenAI key) for text cleanup, extra voices, and language packs. If you have an NVIDIA GPU, it also offers GPU acceleration. You can skip any of this and add it later from Settings.

2. Add a book

1. Open the Library and drag in an EPUB or PDF — or click Add to browse. You can drop several files at once.
2. Other formats (MOBI, AZW3, DOCX, FB2, …) are converted automatically if you have Calibre installed — grab it from the Add-ons page (Download & Install).
3. Confirm the title, author, year, and language in the dialog, then click Import. A fresh install already includes a sample public-domain book to try.

3. Produce the audiobook (Process pipeline)

Open a book and step through the pipeline. Every stage is optional — skip straight to TTS if your text is already clean.

1. AI Cleanup. Fixes OCR errors and formatting, or rewrites text in simpler language for learners. Pick which AI to use (or skip).
2. Translate. Optionally translate the whole book to another language, or build a bilingual, sentence-aligned version for language learning.
3. TTS. Choose the engine, voice, and speaking speed, then generate the narration. On a GPU it runs on the GPU automatically; on a Mac it can use multiple workers.
4. Assembly. Combine the audio into a finished M4B with chapter markers and synced captions (or a video).
5. Review. A summary before you commit. Jobs run in the Queue — start/pause them there and watch progress.

You can set defaults for all of these (AI per step, voice, speed, device, output format) in Settings → Pipeline Defaults so you do not re-pick every time.

4. Listen instantly (live streaming)

1. Open a book and click Start (or start the TTS server from the nav rail). The first time, it loads the voice model into memory — a progress indicator shows when it is ready.
2. Press Play. Audio streams as it is generated; a buffering indicator appears while it gets ahead of you. Adjust speed and voice on the fly, and set bookmarks.

Browser extension (read any web page aloud)

The extension is a Chrome/Edge add-on you load manually — it is not in the Web Store yet, so follow these steps:

1. Download the Browser Extension zip from the button above.
2. Unzip it to a folder you will keep. Chrome loads the extension from this folder, so do not delete or move it afterward.
3. Open chrome://extensions (or edge://extensions) in your browser.
4. Turn on Developer mode (the toggle in the top-right corner).
5. Click Load unpacked and select the folder you unzipped.
6. The BookForge Reader extension appears. Click the puzzle-piece icon in the toolbar and pin it.
7. In the BookForge app, start the TTS server (from the nav rail, or a book’s Start button). The extension connects automatically — no token or login required.
8. On any web page, select text (or use the extension popup) and play it back in BookForge voices.

The BookForge app must be running with its TTS server on for the extension to generate audio.

Pulls your A/B title test results and video stats into a spreadsheet. Takes seconds. Nothing is uploaded anywhere — it saves a CSV to your Downloads folder and that is it. It only reads: it never changes your titles or settings. No revenue or earnings data is collected.

Install

1. Unzip it — and keep the folder. Chrome loads the extension from that folder, so do not delete or move it afterwards.
2. Go to chrome://extensions
3. Turn on Developer mode (toggle, top right)
4. Click Load unpacked and select the unzipped folder

Run it

1. Go to YouTube Studio → Content — the video list. This matters: the extension reads which channel to scan from that page, and will not run from the dashboard.
2. Click the puzzle-piece icon in Chrome’s toolbar, then A/B Title Test Exporter
3. Read the summary and click I understand — continue
4. It will show your channel name. Click Scan selected channels
5. Wait a few seconds. The CSV saves to your Downloads folder automatically and you get a notification when it is done.

If you run more than one channel, open each one’s Content page in its own tab first — they will all be listed, and you can tick whichever you want. They are scanned one after another, each producing its own file.

What you get

One row per video, plus one row per title variant on anything you A/B tested: video ID and link, title, full description, release date, duration, lifetime impressions and click-through rate, views, watch time, average percentage viewed, average view duration, subscribers gained and lost, new versus returning viewers, likes, dislikes, comments and shares — and for each test, every variant’s title, its watch-time share, which one won, and when the test ran.

The winning metric is watch-time share, not click-through rate — that is what YouTube itself uses to decide a title test. Private and unlisted videos are skipped.

How it works: YouTube provides no export for any of this, so the extension reads it from YouTube Studio’s own endpoints using your existing signed-in session. No API key, no password, no account. It only ever reads your own data from your own channel — but automating YouTube Studio is a grey area under YouTube’s terms of service, and the extension says so plainly and asks you to agree before it does anything.