User Guide

Turn any YouTube video into a full, timestamped transcript you can copy, save, or feed to an AI.

1. Paste a YouTube URL (watch, share, or shorts link) and press Get Transcript.
2. Read the transcript with timestamps, or switch to plain text.
3. Copy it, or use Save TXT / Save SRT to choose a filename and folder in supported browsers.

• 💡No account is needed for occasional use; sign in for history and higher limits.
• 💡Videos with captions disabled by the uploader can't be transcribed.

Extract the text from PDF, DOCX, PPTX, XLSX, TXT, MD, EPUB, and HTML files — per page, section, slide, sheet, or chapter.

Read the full Documents guide →

1. Drag a file onto the dropzone (or click to browse), or paste a public link to the document. Files up to 25 MB are accepted (200 MB on Pro and API plans).
2. Browse the extracted text by page, heading, or slide.
3. For PDFs, open the Markdown tab for a clean, structured version with page anchors.
4. Download the result as TXT, JSON, or Markdown — or, for a PDF or EPUB, press Open in PDF Editor to edit, sign, watermark, or password-protect the same file.
5. On Pro and API, translate a page or the whole file into 30 languages under the result and download it as DOCX, PDF, or TXT.
6. On Pro and API plans, drop up to 25 files at once for a batch and download them as one ZIP or merged file, and press Download chunks (JSON) for heading-aware chunks ready for a vector database.

• 💡Scanned PDFs and images (PNG, JPG, TIFF, WEBP, BMP) are read with OCR on Pro and API plans (3 units a page, up to 50 pages); on Free they report that OCR is a Pro feature.
• 💡Re-uploading the same file is instant: results are cached by content hash.

Annotate, sign, merge, split, and reorganize PDF files — processed in memory, never stored.

Read the full Manage Pages guide →

1. Upload a PDF — or an EPUB, which opens as a text-only PDF — to edit it with text boxes, shapes, signatures, images, notes, whiteouts, or a watermark across every page.
2. Open Manage Pages to organize, split, or merge PDF files, add a password to a PDF, or unlock one you have the password for — without leaving the editor.
3. Save a flattened edited copy, print it straight from the editor, or download the processed merge, split, or organized result.

Collect every comment and reply on a YouTube video, score them with AI for sentiment, emotions, toxicity, and language, and export the table as CSV, Excel, or JSON.

1. Open Sentiment Analysis from the paid tools section.
2. Paste a YouTube link, or switch to Batch for up to ten links or a playlist. Pick Top comments or Newest and how many to collect (replies included).
3. Tick Analyze with AI to score the comments as they arrive, or press Analyze with AI on a collected set. Choose how many comments the pass reads; top-level comments go first.
4. Read the summary (sentiment split, average and like-weighted score, emotions, toxic share, languages, strongest comments each way), filter and sort the table, then save it as CSV, Excel, or JSON.

• 💡Sentiment Analysis requires Pro or API. Collection costs 1 unit per started 100 comments (a stored repeat 1 unit); the AI pass costs 10 units per started 100 comments analysed. Batches meter each video that succeeds.
• 💡Pinned, verified, uploader, and creator-heart flags come from YouTube's page data; when the server collects through the YouTube Data API those show as unknown.
• 💡Toxic comments are flagged at a toxicity of 0.5 and above; the Toxic filter lists them for review.

Extract public pages into clean Markdown, text, HTML, JSON, or structured fields, then summarize and ask questions about them in your Library.

1. Open Web Scraping from the paid tools section.
2. Paste one public page URL, or switch to Batch and paste up to 25 URLs, one per line. In Batch, "Discover pages" lists a site's pages from its sitemap and links so you can tick the ones you want.
3. Pick a fetch method in Options: Plain, Auto (plain first, browser only if the site refuses), or Always render for JavaScript apps. Add CSS selectors for structured fields, and set wait, device, screenshot, and PDF options when a browser is available.
4. Review the result across the Markdown, Text, HTML, JSON, Links, and Data tabs, then copy, save, screenshot, print to PDF, or open the page in your Library.

• 💡Web Scraping requires Pro or API. A plain page costs one unit; a browser-rendered page, screenshot, or PDF costs three; Auto bills whichever path served the page; discovering a site's pages costs one unit. Batches meter successful pages only.
• 💡Private-network targets are blocked, redirects are revalidated, and site robots guidance is respected. Bot-protection bypass, CAPTCHA solving, and logged-in sessions are deliberately not offered.
• 💡Extracted pages are cached for an hour and appear in your History as Web sources with summaries, Ask, and Create From Source.

Search open collections for books, papers, magazines, and comics, read the catalog record, and download what is free to keep.

1. Open Open Library from the paid tools section.
2. Type a title, author, or subject (or isbn:… / doi:… to look one item up). All sources, the default, searches every source at once, merges records of the same work, and ranks free copies first; use the chips to narrow by source, kind, access, format, or language, or pick one source: Books on Google Books (the broadest catalog), Open Library, or Project Gutenberg; Papers on arXiv or the Internet Archive; Scholarly records on OpenAlex; Magazines, Comics, and Everything on the Internet Archive.
3. Each result shows its access level. For items marked free to download, press Show downloads and pick PDF, EPUB, text, or comic-book formats.

• 💡Searching is free on Pro; each download uses one unit and is capped at 300 MB.
• 💡Downloads open a Save As dialog on the click (Chrome/Edge): pick the folder, keep or edit the suggested name — the book's title — and the file streams straight there; cancelling costs nothing. Other browsers download the titled file to their downloads folder.
• 💡In-copyright titles never download here; Google Books results offer Preview, Borrow (Open Library), Find in a library (WorldCat), and Buy links instead, and paywalled papers link to the publisher.

A Pro workspace for turning one recording into many assets: download or upload a video, transcribe it, find the best moments, then trim, caption, crop, and compress them for each platform. Every tab has its own link (?tool=…), and one tab's result can feed the next.

Read the full Media Tools guide →

1. Upgrade to Pro (or use an API-plan account), then open a tab from the list on the page or through its link, for example /media-tools?tool=crop.
2. Tabs that work on a file share one uploaded video: load it once and Trim, Captions, Compress, and Crop all use it, along with the preview player's in/out points.
3. Tabs that work on a link (Download, Social) or a recording (Audio Kit, the transcript tabs) take their own input.
4. Results download immediately and are not stored, except transcripts, which are kept so repeats are cheaper and so Clip Ideas can use them.

• 💡The MP3 tab exports the audio of the video loaded in the preview; Download's Audio only field does the same for a public link.
• 💡Play YouTube previews a video in the player for scrubbing and marking ranges; exports still need the file, so Download it first.
• 💡Pricing in one sentence: 1 unit per file export, 1 unit per MB for public downloads, transcription per minute (minimum 3), AI work 10 units (30 at Deep quality), and exact repeats 1.
• 💡The old /audio and /video pages open the matching transcript tab here.

Summarize any transcript or document you've extracted: overview, key points, and chapters — timestamped for videos, page-anchored for documents.

1. Sign in, then extract a video or document first (the summary works on a saved extraction).
2. Open AI Summary (or press Summarize on a result) and choose the source.
3. Read the overview, key points, and chapters; chapter links jump straight to the timestamp or page.

• 💡A fresh summary costs 10 quota units at Standard quality or 30 at Deep, which runs on the largest model; repeating one that's cached costs 1. Pick Standard or Deep next to the Summarize button — the choice is remembered.
• 💡AI summaries, Ask and Create are Pro and API features: they run on Claude, take any length, and give the full summary with chapters, action items, quotes and questions. A Free account sees the standard slice (overview, key points, topics) of summaries it made before, and the Pro note in place of the AI panels.

Listen to any text read aloud with a female or male voice, at your own speed, with the spoken word highlighted live.

1. Paste or type your text.
2. Pick Female or Male, choose a specific voice if you like, and set the speed.
3. Press Play for free browser playback, or use Download MP3 on a Pro or API plan for a neural voice file.

• 💡Voices come from your browser and operating system, so the selection differs per device.
• 💡Browser playback is free and local; neural MP3 export runs on a hosted voice provider, so it is a Pro feature and uses quota.

Dictate live with your microphone, or upload a recording (meeting, podcast, voice memo) and let Whisper transcribe it.

1. Dictate: choose your language, press Start dictating, and speak — interim words appear in grey and settle into the transcript.
2. Upload: sign in, pick a recording (any common audio or video format, up to 500 MB), and press Transcribe. 1 unit per started minute, minimum 3; cached repeats 1.
3. Edit the text directly, then copy it or download it as a TXT file.

• 💡Live dictation works in Chrome, Edge, and Safari; on Firefox use the Upload tab instead.
• 💡Uploads cost 1 unit per started minute of audio, minimum 3 (repeating the same file costs 1); dictation is free and unmetered.

Write, edit, and preview Markdown live — or open a document extraction and polish it before sharing or feeding it to your AI.

1. Type in the left pane and watch the preview update live — or switch to Edit or Preview only.
2. Use the toolbar (or Cmd/Ctrl+B, I, K) for bold, headings, links, lists, quotes, code, and tables.
3. Press Img to open a folder and search its pictures by name, pick files, or link to an image on the web. Add picture, paste and drag-drop embed one directly. Pictures show in the preview and travel inside the saved file; while you edit, each is a short embedded:name.jpg stand-in.
4. Press Save as… to pick a folder and name (Chrome/Edge open a native dialog; other browsers download using the file-name field), or Print to send the rendered preview to your printer.

• 💡Runs entirely in your browser — free, no account, and your draft autosaves locally.
• 💡Pictures are embedded into the Markdown itself as data URIs, so the saved .md is self-contained.
• 💡Extractions keep their page/section anchors as HTML comments — harmless in the preview, useful for AI chapters.

Sign in for history, AI, and API keys: Free has 100 units, Pro 2,000, and API 10,000 plus paid tools.

1. Register with your email, then manage everything from the Account page.
2. Mint an API key (shown once — store it safely) to call the extraction endpoints programmatically.
3. Track this month's usage, and upgrade to Pro or API via secure Stripe checkout when you need more.

• 💡API calls authenticate with “Authorization: Bearer tf_live_…” and return X-Quota headers so you always know where you stand.
• 💡History shows every past extraction so you never fetch the same thing twice.