Blinded

Blinded: pricing, features and limits

Zero-upload redaction that finds every copy of a name, logo or face by sight

Blinded costs from $2.99 USD one-time, as of 29 September 2026.

Key facts

Vendor
Blinded
Headquarters
Not recorded
Media
image, pdf
Deployment
online
Method
manual-redaction, rule-based
Free tier
Yes
Trial
None
Entry price
$2.99 one-time

What it does

  • OCR
  • True content removal

Screenshots

Plans and pricing

Blinded plans. Every figure records where it came from.
PlanPriceIncludes
FreeFreeDocuments up to 20 pages · No account needed · Full feature set, zero-upload
  • Pages per document: 20 pages
5-Day License$2.99 one-timeUnlocks export for documents over 20 pages · Valid for 5 days · No account — a standalone license string
  • File retention: 5 days
1-Month License$6.99 one-timeUnlocks export for documents over 20 pages · Valid for 1 month · No account — a standalone license string
  • File retention: 30 days

read automatically from the vendor · entered by our editors · supplied by the vendor and not independently verified. Check the vendor's pricing page.

What will it cost?

100 pages a month

Worked out from each plan's published rate, reading the figures above as one month's work. Subscription plans read Included in the plan rather than a per-document figure: a monthly fee is not a per-document price, and dividing one into the other would need a volume no vendor publishes. A metered plan is priced past its allowance at its published overage rate; one that caps out with no such rate says so instead of quoting a price it would not honour.
PlanPublished priceTotalPer page
FreeFree—
5-Day License$2.99 one-timeIncluded in the plan
1-Month License$6.99 one-timeIncluded in the plan

Our assessment

Blinded is a client-side-only redaction tool built around a single claim: nothing is ever uploaded. The document is opened, read and rewritten entirely inside the browser tab — no server round-trip, no account, no analytics, and explicitly no LLM used for detection. The vendor invites users to verify this themselves via a dedicated verification page and publishes the source on GitHub.

Detection works by visual matching rather than AI: pick a name, logo, stamp or image once, and every occurrence is found and removed regardless of size or colour, including instances baked into a scanned page or inside artwork/photos rather than selectable text. Because the page is rebuilt from pixels, redaction is a true removal rather than a visual overlay — the tool's own comparison table contrasts this with Adobe Acrobat Pro (where a box must be drawn and then separately "Sanitized" to actually remove content). Redactions can optionally be labelled (e.g. [P1], [E2]) so a later human reader — or a model the document is later pasted into — can still follow the sentence structure around what was removed.

It targets one-off, high-sensitivity document handling rather than ongoing programmatic redaction: M&A deal materials (CIMs, board decks, diligence packs) where a counterparty's name and logo must come out of every slide including small logos inside charts, legal agreements before sharing as precedent, FOIA/discovery productions, de-identifying medical records for research or teaching, blind CV screening, and stripping names/logos out of a document before pasting it into an AI chat tool.

Supported inputs are PDF, PNG, JPEG, and the tool's own saved-draft format, with a companion set of ordinary PDF utilities (extract, rotate, add text, save a draft and resume later); it also works on mobile browsers. Documents up to 20 pages are entirely free with no account. Beyond 20 pages, a license is required to export: a short-lived license costs $2.99 for five days, or $6.99 for a month — everything up to the final export/save step (opening, searching, marking, reviewing, saving a draft) stays free at any document length. Licenses are standalone strings rather than accounts: nothing to sign up for or cancel, and they simply stop working on expiry.

Strengths

  • True zero-upload: processing happens entirely client-side in the browser tab, with a page inviting users to verify this and public source code
  • Finds recurring names, logos and images by sight and removes every copy in one pass, including inside scanned pages and artwork — not just selectable text
  • True pixel-level removal rather than a visual overlay that can be undone or selected through
  • Optional labelled redactions (e.g. [P1]) keep surrounding sentences readable to a later human or model
  • Genuinely free for documents up to 20 pages with no account required, and cheap one-off access beyond that ($2.99/5 days or $6.99/month) rather than a mandatory subscription

Limitations

  • No AI/LLM-based entity detection — relies on the user picking each name/logo/image to match, so nothing is found automatically without a first selection
  • 20-page free cap means most real deal decks, discovery productions or medical charts require a paid license to export
  • No stated compliance certifications (HIPAA, SOC 2, etc.) — privacy relies on the client-side architecture rather than a certified program
  • Licenses are simple time-boxed strings with no account, so there's no team/seat management, audit log or centralized administration for organizational use
  • Vendor explicitly disclaims certification: output is not certified and redaction mistakes may occur, so manual review is required before relying on it
  • Rule based search works bad on our benchmarks.

Do you work for Blinded?

Claim this listing to propose corrections to its pricing and details.