Pick the source
Drop a file, paste captions, or reference a link for attribution. Nothing starts until you confirm you have the rights to it.
NewYour video now stays on your device
Turn any video into an article you can verify.
Video2Blog reads what is said and what is shown — slides, product screens, numbers — and writes a structured article where every paragraph links to its exact moment. You review the evidence, not a wall of AI text.
Free · 20 minutes of video · no card · runs on Cloudflare Workers AI
Source 23:48
0:0011:5423:48
Evidence 3 moments
04:17
Plans compared
ON SCREEN Starter · Growth · Scale
10:57
Free → Paid 40%
ON SCREEN Conversion by plan — 40%
18:34
Annual −20%
ON SCREEN Annual billing −20%
Article draft
Why usage-based pricing converts
The team compared three plans side by side before changing anything. ▶ 4:17
After the switch, the free plan converted at 40% once customers paid per processed minute. ▶ 10:57
inferenceThe free tier now works as a trial for heavy users. ▶ 11:20
92
Quality 92 · 0 uncited claims · quotes verifiedYour video file stays on your device
Every paragraph cites its moment
Runs on Cloudflare Workers AI
Claims checked without AI
How it works
From recording to reviewed article, in seven steps.
The heavy lifting on the video happens in your browser. The model work happens on Cloudflare. Every step leaves evidence you can check.
- Your videoMP4 · MOV · WebM · audio
Drop a file, paste captions, or reference a link for attribution. Nothing starts until you confirm you have the rights to it.
- Local prepin your browser
The file is decoded on your device. It is never uploaded or stored — only what the next steps need leaves your machine.
- Audio · frames · OCRin your browser
Audio becomes 30-second parts. Key frames are picked where the screen changes, and their text is read by OCR — slides, UI labels, numbers.
- TranscriptionWorkers AI · Whisper
Each audio part is transcribed once with timestamps and language detection. A part is never transcribed — or paid for — twice.
- AI writingWorkers AI · Gemma
Speech and screens are merged into ~45-second evidence windows, then written into an article in one structured call. Every paragraph must cite its windows.
- Articlechecked without AI
A deterministic pass checks citations, verbatim quotes and numbers against the source, and points at the exact paragraph to review.
- Derived contentLinkedIn · X · newsletter
From the reviewed article: LinkedIn post, X thread, newsletter section, YouTube description and quotes — same citations, one call.
Product demo
See exactly what you get.
Click a transcript line or a key moment: the video, the timeline and the article stay in sync. This is sample data — your own video gives you the same views.
Interactive demo · sample data
Transcript · click a line
Why usage-based pricing converts
What one SaaS team learned by dropping per-seat plans
The team put three plans side by side before changing anything; the cheapest one mostly served as an anchor.
Free → Paid conversion40%
The number that surprised them
“On the free plan we now see a forty percent conversion rate” — once customers paid per processed minute instead of per seat.
inferenceThe free tier now behaves like a trial for heavy users.
We stopped charging per seat. Our free plan now converts at 40% — here is what changed, with the exact moments from the webinar. ▶ 10:52
1/ Per-seat pricing hid our best users. We switched to usage. The free → paid conversion went to 40%. 🧵
This week: why usage-based pricing converted 40% of free users — and why the team treats the free plan as a trial for heavy users.
Try it on your video Takes a few minutes for a 20-minute webinar.
Product
Everything between the recording and the published post.
One evidence timeline
Speech, key frames and on-screen text aligned in ~45-second windows. The article, the checks and the search all read from it.
Key frames that matter
Picked where the screen changes — a duplicate only if three signals agree.
What was shown, read
Slides, UI labels and numbers, with a confidence score. Low-confidence reads are flagged.
Checked without AI
Citations, verbatim quotes, numbers absent from the source — computed on every save.
Timestamp citations
Review the evidence next to the claim instead of re-watching an hour.
A library you can search
Find what was said or shown across every source you processed.
Repurpose & export
Derived formats keep their citations. Identical requests reuse what was already generated — no second bill.
Your video filedirectCaptions / .vtt / .srtdirectAudio filedirectZoom & Meet recordingsdirectScreen recordingsdirectYouTube (link + your file)file neededVimeo (link + your file)file neededReels & TikToks (your file)file neededLoom exportsdirectCourse lessonsdirect
The difference
Same webinar. Two very different drafts.
Transcript-only tools summarise what they heard and fill the gaps. Video2Blog writes only what the source says or shows — and proves it.
Transcript-only tool
Usage-based pricing is a proven strategy used by most successful SaaS companies. The team saw conversions roughly double after switching.
- Fills gaps with plausible filler
- Misses what was on screen
- No way to check a claim
Video2Blog
After the switch, the free plan converted at 40% once customers paid per processed minute. ▶ 10:52
inferenceThe free tier now works as a trial for heavy users. ▶ 14:50
- Only what the source says or shows
- Slide numbers captured by OCR
- Inference labelled, every claim linked
Privacy & architecture
Built so your video never leaves your device.
No black box: here is where each step runs and what is kept.
Your browser
Your file stays with you
- Video decoded locally (WebCodecs / Web Audio)
- Key frames chosen on-device
- On-screen text read on-device (Tesseract, self-hosted)
- Only 30-s audio parts + up to 60 screenshots are sent
Cloudflare Workers AI
One model call per job step
- Whisper large-v3-turbo transcribes each part once
- Gemma writes the article in one structured call
- Results cached by input: retries never pay twice
- No third-party AI provider, no keys to manage
Your workspace
Evidence you can check
- Transcript, screenshots and on-screen text kept for editing
- Audio parts discarded after transcription
- Export everything as JSON, delete anytime
- Content is never used to train models
30 saudio parts, each transcribed exactly once
60key frames at most — chosen, not sampled
1AI call to write an article (formats: one more)
0AI calls in the quality check
Pricing
Pay for video hours. Edit freely.
No credits to decode. Re-writing, AI edits, language versions and exports of a processed source don't count.
Free
$0
Try it on one video.
- 20 minutes of video / month
- Up to 20 min per source
- All exports & edits
Creator
$19 / month
For a weekly video or podcast.
- 5 hours of video / month
- Up to 2 h per source
- Brand profile · batches of 10
Pro Most teams
$49 / month
For teams publishing from demos and webinars.
- 20 hours of video / month
- Up to 3 h per source
- Batches of 25 · 4 jobs in parallel
Team
$129 / month
For agencies and content teams.
- 60 hours of video / month
- Up to 3 h per source
- Batches of 50 · 8 jobs in parallel
Prices in USD. Paid plans are opening in stages — during the preview, contact us to upgrade. Unused hours don't roll over.
FAQ
Questions, answered plainly.
Is my video uploaded to your servers?
No. When you add a file, your browser decodes it locally. Only 30-second audio parts (for transcription) and up to 60 selected screenshots are sent. The audio parts are discarded once transcribed; the transcript, screenshots and on-screen text are kept so you can edit and re-generate.
Which AI does Video2Blog use?
Cloudflare Workers AI: Whisper large-v3-turbo for speech and Gemma for writing. There is no third-party AI provider and no API key to bring. Your content is not used to train models.
How do you prevent invented facts?
The writer only sees the evidence timeline and must cite it. A separate, non-AI check then verifies that citations exist, quotes appear word-for-word in the transcript and numbers appear in the speech or on screen. Anything that fails is flagged at the exact paragraph.
Can I paste any YouTube link?
You can paste it to reference the video: we read its public title, channel, thumbnail and chapters for attribution and timestamp links. YouTube's terms don't allow third-party services to download videos, so to build the article we need the file (if it is yours or you have permission) or its captions — both export from YouTube Studio.
Can I import an Instagram or TikTok account by @username?
No. Neither platform offers an approved way for a service to list another account's videos, and we don't scrape them. Upload the videos you have rights to; paste the post link to keep it as the credited source.
Which sources work?
Video and audio files (MP4, MOV, WebM, MP3, M4A, WAV) prepared in your browser, and captions or timestamped transcripts (.vtt, .srt, .txt). Direct media links and podcast feeds are fetched server-side on deployments where the media engine is enabled.
Start with one video
Your next article is already recorded.
Drop in a demo, a webinar or a tutorial. Review the evidence, not a wall of AI text.
Free · 20 minutes of video · no card