Disclosure up front: I make one of the tools on this list (Outbound Veil, a Mac menu-bar app). In several cases a competitor is the better choice. Predact is free and handles files, which Outbound Veil does not. PrivacyScrubber works on Windows and Linux, which Outbound Veil does not. If your data cannot leave your machine at all, a local model beats every redaction tool here.
Prices and platforms were read on each vendor's own page on October 4, 2026, and they change. MetaVeils was added on October 9, 2026, and I read its page that day. Data Guard, Prompt Guard and Whiteout were added on October 14, 2026, each from a page I read that day. Redaktr and MaskPaste were added on October 21, 2026, the same way. Where a vendor says its software runs on your device, I write "says": I have not audited that.
The short answer
Start with your chatbot's own account controls, which decide whether chats are used for training. A developer scrubbing transcripts can use Presidio, a free library. A browser extension such as PrivacyScrubber suits several AI sites in Chrome. A Mac menu-bar checker such as Outbound Veil, which I make, flags details as you type. Prices read October 4, 2026.
If the data must never leave your machine, a local model beats every redaction tool here. No tool listed makes a workflow compliant, and all of them miss some things.
The tools side by side
"Not stated" means the guide and the vendor pages it cites do not say. Where detection happens is the vendor's claim unless the row says otherwise.
| Tool | Price | Runs where | Where detection happens | What it covers | Notable limit |
|---|---|---|---|---|---|
| Outbound Veil (mine) | $29.99 once, after a free 7-day trial | macOS 14 or later, Apple silicon and Intel | On your Mac: a model that runs locally does the checking, and nothing you type is stored or sent | Text in the field you are typing in: names, addresses, phone numbers, email addresses, government, tax, bank and card numbers, dates of birth, medical record numbers, IBANs | Mac only, names in Latin script only, checks text, not files, will miss some things, not a compliance product |
| Predact | Free, you enter an email address to get the download link | Mac (macOS 14 recommended) and Windows | Its page says processing is local, Safe Mode goes online only when you export to an AI service, and Fully Offline Mode blocks all network use | PDFs, Word files, text files, images, email clippings, pasted text and screenshots | Does not check what you type in other apps, the local-processing statement is the vendor's claim |
| PrivacyScrubber | Free up to 15,000 characters per scrub, Pro $15 a month or $110 once, Teams $99 a month, developer SDK $299 a month | Web app, Chrome extension, MCP server and Node.js SDK | Its page says in the browser tab with no server contact, I have not tested that | Chat boxes on 12 listed sites, including ChatGPT, Claude, Gemini, Copilot and Perplexity, with the original values restored in the reply | Browser only, you press the button, larger inputs, PDF OCR and custom rules need a paid plan |
| Microsoft Presidio | Free, MIT license | Python library, with Docker and Kubernetes options | Wherever you run it, with no external service | Personal data in text, images and structured data | No interface and no guarantee it finds everything, you write the code that calls it |
| The provider's own controls | Settings free, business plans priced by OpenAI | Any | Not detection: your text still reaches the provider's servers | Whether chats are used for training, nothing is redacted | Text is kept for some period, Temporary Chat may keep a copy for up to 30 days for safety |
| A local model (Ollama, LM Studio) | Free to download | macOS, Windows and Linux | On your computer, Ollama's :cloud models run on Ollama's servers instead | Whatever you paste, no redaction is needed | Setup time and memory, open models that fit on a laptop are generally less capable than the largest hosted ones (my judgment) |
| Find and replace by hand | Free | Any text editor | Nowhere, you do it yourself | Whatever you read and replace | Unreliable on long text: a misspelled surname, a nickname or an employer slips through |
| Caviard | Free for up to 10 protected values per document, lifetime plan $9, listed as an early discount from $15.30 | Chrome extension for ChatGPT and DeepSeek | In the browser, before a prompt is sent | Pattern-based detection, which suits emails and card numbers | Tends to be weaker for names, I have not tested it on names |
| ClipScrub | $29 once, 14-day free trial | macOS 15 or later, AI-enhanced detection needs Apple silicon | On the Mac with no internet needed, according to its page | PDFs, Word files, screenshots, screen recordings, JSON, CSV, XML and clipboard text | Does not watch your typing, says it is not a compliance service and does not guarantee complete de-identification |
| Clipmask | Free to download, its page also mentions a 500-copy trial limit and gives no price for continuing | Apple silicon Macs only, macOS 14 or later | Its page says on the Mac, the only server contact is a check that the app is a licensed current version | Text you copy, Option+Space offers a redacted version to paste | Clipboard only, Apple silicon only |
| RedactDesk | Free, open source (MIT) | macOS 14 or later | Locally, with OpenAI's open-source privacy-filter model | PDFs, scanned ones included, it deletes the text rather than covering it | PDFs only |
| BlurData | $79 a year or $149 once for one person, team plans listed, no free tier shown | macOS 13 or later, with an iOS version | On the device | JPG and PNG screenshots and PDFs, in batches | Screenshots and PDFs, not typed text |
| MetaVeils | Free edition, Pro Lifetime License $39.99 once | macOS, as a .dmg, minimum macOS version not stated | Its page says 100% offline, with scanning on your Mac, not tested | TXT, PDF, DOCX and images, names, emails, phone numbers, addresses, card and bank numbers, Social Security numbers, GPS coordinates, metadata | Free edition limits not stated, image OCR and metadata stripping need Pro, whether it checks clipboard text or typing is not stated |
| Nightfall | By quote, per user per year, in two editions | Browser, endpoint, SaaS and AI-app coverage, run by an administrator | Not stated | Policies that block or flag sensitive data going into AI apps, email and other services | Built for organizations, a security team has to set the policies |
| Microsoft Purview | Depends on your Microsoft licensing, which I did not confirm | Managed devices, Edge and other browsers | Not stated | Endpoint DLP policies that block pasting and uploading sensitive information to AI app websites, Edge policies that inspect prompts in consumer AI apps | Needs IT to turn it on for managed devices |
| ChatGPT Privacy Shield | Free Chrome extension, Pro ($9 a month) and Team ($49 a month) listed as a waitlist | Chrome extension, ChatGPT only | Not stated | Not stated | ChatGPT only, a young project |
| Rescriber | Free, open source | ChatGPT extension from Northeastern University researchers | Not stated | Not stated | ChatGPT only, a young project |
| PII Scrubber | Free for 5 uses a day, Pro $7 a month | Extension for ChatGPT | Not stated | Not stated | Small extension with paid tiers |
| Paste Redactor | Free for 100 redactions a month, then $1 a month | Extension for four browsers | Not stated | Not stated | Small extension with paid tiers |
| PIIGuardAI | Open source, price not stated | Mac menu-bar proxy | Not stated | Blocks AI requests containing patterns such as emails and card numbers | Uses regular expressions only |
| Maskbase | Open source, price not stated | Mac app | Not stated | Documents | A young project |
| Veil (helloveil.com) | $8 a month | Not stated | Not stated | A meeting recorder with a privacy engine | A different company's product, its changelog lists nothing that checks what you type in other apps |
| Data Guard (Firefox) | Not stated on the listing, as listed on 2026-10-14 | Firefox add-on, for chatgpt.com, chat.openai.com, claude.ai and gemini.google.com | Not stated | Its listing says it warns you before you share Social Security numbers, card numbers, API keys, phone numbers, street addresses, passwords and health-related information | Firefox only. The listing says the warning does not block or change your message. As listed on 2026-10-14 it showed version 1.0, an update on July 31, 2026, one review and no users. Not tested by us |
| Prompt Guard (Firefox) | Not stated on the listing, as listed on 2026-10-14 | Firefox add-on, for chatgpt.com, gemini.google.com, claude.ai and a few other AI sites named in its permissions | Its listing calls it client-side | Its listing says it detects and masks personal information, API keys and credentials as you type or paste, with built-in rules for card numbers, Social Security numbers, API keys and passwords, and custom rules you can add | Firefox only, and a young project: as listed on 2026-10-14 it showed version 0.1.0, an update on May 25, 2026, one user and no ratings. Not tested by us |
| Whiteout (Mac) | $9.99 once, with a 7-day free trial, as listed on hunted.space on 2026-10-14 | Mac menu-bar app | The listing says fully offline, with no cloud, no accounts and no data leaving your machine. That is the listing's claim | Screenshots: the listing says it scans them for details such as phone numbers, Social Security numbers, credit cards and API keys, saves a redacted copy and keeps the original in a backup folder | Screenshots only, not typed text. I could not find the developer's own page, so this row rests on the hunted.space listing alone, as listed on 2026-10-14. Not tested by us |
| Redaktr (Mac) | Free trial, then $39 once, as listed on Product Hunt on 2026-10-21 | Mac app | The listing says scanning and redaction run entirely on the Mac with no server involved, and that license activation and update checks do use the network. That is the maker's claim | Screenshots: the listing says it checks them for secrets, customer data, faces, QR codes and policy terms before you share them | Screenshots only, not typed text. As listed on 2026-10-21 and not tested by us |
| MaskPaste (Mac) | Not stated on the listing, as listed on 2026-10-21 | Mac app | The listing calls it local-first. That is the listing's claim | Text: the listing says you select names, company names and addresses, choose replacements, copy the masked text into an AI tool, then paste the reply back to restore your originals | Its only listing I could read is on hunted.space. As listed on 2026-10-21 and not tested by us |
| SinapticAI (Mac App Store) | Free, as listed on 2026-11-04 | Browser extension, delivered through the Mac App Store | The listing says analysis happens on your device. That is the listing's claim | Pasted text: the listing says it warns when you are about to paste personal, financial or private-key data into a website, with rules to block, warn or audit | Paste only, in the browser. As listed on 2026-11-04 it showed version 1.2 from January 7, 2026 and no ratings. Not tested by us |
Sources. Rows come from the sections below and the vendor pages they cite, read October 4, 2026. MetaVeils: metaveils.com, read October 9, 2026. Not tested. The last eight rows are the narrower options described near the end of the tool sections. Data Guard: its Firefox add-on listing, read October 14, 2026. Prompt Guard: its Firefox add-on listing, read October 14, 2026. Whiteout: its hunted.space listing, read October 14, 2026. Those three rows are as listed on 2026-10-14 and not tested by us. Redaktr: its Product Hunt listing, read October 21, 2026. MaskPaste: its hunted.space listing, read October 21, 2026. Those two rows are as listed on 2026-10-21 and not tested by us. SinapticAI: its Mac App Store listing, read November 4, 2026, as listed that day and not tested by us.
What you actually need to stop leaking data into AI chats
Redaction means replacing specific details in your text (such as a name, a date of birth or a card number) with a placeholder before the text goes to the AI. The model still answers, because it sees "[PERSON_1] was seen on [DATE_1]" and can work with that. You put the real details back yourself, or a tool does it from a saved mapping.
There are four places to intervene, and the tools here sit in different ones:
- As you type or paste. A menu-bar app or browser extension watches the text box and flags or replaces details before you press send. This is where the mistake usually happens.
- Before you paste. An app or web page where you drop in text or a file and paste the clean copy. It works on the days you remember it.
- At the account. The provider's settings and plan decide whether what you send is used for training. This removes nothing from their servers.
- Off the cloud. A model that runs on your own computer, so nothing is sent.
All of them share two limits, covered in what redaction cannot do. Detection misses things, and a redacted note can still describe a situation someone could recognize.
The shortlist
- Outbound Veil: $29.99 once, a Mac menu-bar app that checks the text field you are typing in.
- Predact: free, Mac and Windows, drop in documents, images or text.
- PrivacyScrubber: browser-based, free tier, extension for 12 chat and work sites.
- Microsoft Presidio: free open-source library for developers.
- The provider's own controls: ChatGPT settings, Temporary Chat, business plans and the API.
- A local model: Ollama or LM Studio, so nothing leaves your machine.
- Find and replace by hand: free, slow, and still useful.
Those seven cover most situations. Eight more tools and eight narrower ones follow.
Outbound Veil
Cost: $29.99, one-time, updates included, after a free 7-day trial with no account. Platform: macOS 14 or later, Apple silicon and Intel. Mac only. Status: version 1.1.0.
Outbound Veil lives in the menu bar and reads the text field that has focus, through macOS Accessibility. When it finds something it shows a badge with the number of findings and highlights the matching text where the app allows it. You can ignore a finding, or press Redact all, which rewrites the field in place with labelled placeholders such as [SSN_1] or [EMAIL_1]. In apps you choose, it can also hold the send key until you have decided.
It looks for names, addresses, phone numbers, email addresses, government, tax, bank and card numbers, dates of birth, medical record numbers and IBANs.
A model that runs on your Mac does the checking. Nothing you type is stored or sent, and the app has no analytics. The model is Rampart by National Design Studio, released under CC BY 4.0, and I did not train it. Purplelink ported Rampart's processing code to Swift and added rules for dates of birth, medical record numbers and IBANs. The model file is unchanged.
What it's good at: sitting where the mistake happens, in any Mac app it can read rather than one website. I tested TextEdit, Notes, Microsoft Word, Messages and text areas in Chrome. In the ChatGPT and claude.ai prompt boxes in Chrome, Redact all works, but the whole box is outlined instead of each word highlighted.
What it lacks: Mac only, and names in Latin script only. It flags the names of public figures too, which you can ignore, and it will miss some things and flag some things that are not personal information. It checks text in fields, not files, so a PDF or an image needs another tool from this list.
Some apps do not expose their text fields to macOS. The Apple Mail compose window was the hard case: since version 1.3.0 Outbound Veil reads it and highlights details there, but you edit them by hand. Safari, Slack, VS Code and the ChatGPT and Claude desktop apps are not on the tested list (the full table). It needs the Accessibility permission, plus Input Monitoring if you turn on send hold.
After Redact all it does not put the original values back into the AI's answer. You swap them back yourself. By default it does not report a bare city, state or ZIP code. It is new, with no track record, and it is not a compliance product.
Pick Outbound Veil if you write or paste client, patient or student details into chat boxes on a Mac and want a check as you send. Pick something else if you are on Windows, mostly share files or screenshots, or your data has names in non-Latin scripts. The trial is free for seven days. Keeping it is $29.99, once.
Predact
Cost: Free, no card. You enter an email address to get the download link. Platform: Mac (macOS 14 recommended, Apple silicon supported) and Windows.
Predact is a desktop app where you drop in PDFs, Word files, text files, images and email clippings, or paste text and screenshots. It shows what it found and lets you select more words to redact. You then copy the result out or open a built-in browser linked to ChatGPT, Claude, Gemini or Perplexity.
Its page says processing is local. A Safe Mode goes online only when you export to an AI service, and a Fully Offline Mode blocks all network use. "Persona filtering" swaps in consistent made-up details instead of tags.
What it's good at: being free, handling files and images, and running on Mac and Windows. What it lacks: it does not check what you type in other apps, so using it is a step you have to remember. The local-processing statement is the vendor's claim.
Pick Predact if your risk is documents, scans or screenshots, you are on Windows, or the budget is zero. Outbound Veil's case is that the check happens while you type, whether or not you remembered.
PrivacyScrubber
Cost: Free up to 15,000 characters per scrub. Pro is $15 a month or $110 once. Teams is $99 a month. A developer SDK is $299 a month. Platform: a web app, a Chrome extension, an MCP server and a Node.js SDK.
The Chrome extension adds a shield button to chat boxes on 12 listed sites, including ChatGPT, Claude, Gemini, Copilot and Perplexity. Click it, or press Alt+Shift+X, and the text becomes placeholders such as [NAME_1]. After the answer arrives, "Reveal Original Data" restores the real values inside the page. Its page says all of this runs in the browser tab with no server contact.
What it's good at: the full loop of redact, ask, restore, on many sites, on any system that runs Chrome. What it lacks: it lives in the browser, so it does not help in a desktop app. You press the button, so a rushed send can skip it. Larger inputs, PDF OCR and custom rules need a paid plan. I have not tested the local-processing claim.
Pick PrivacyScrubber if you use several AI sites in a browser, want the original values restored in the reply, and either use Windows or Linux or do not want an app.
Microsoft Presidio
Cost: Free, MIT license. Platform: a Python library, with Docker and Kubernetes options.
Presidio is an open-source framework for finding and anonymizing personal data in text, images and structured data, using named-entity recognition, regular expressions, rules and checksums. You can write your own recognizers, and it runs wherever you run it with no external service. The project warns there is no guarantee it finds everything. It has moved from Microsoft's GitHub organization to a community-owned one, and its documentation says that transition is under way.
What it's good at: control. You can run it over a folder of interview transcripts and keep the pipeline in your own infrastructure. What it lacks: an interface. It will not watch a chat box, and nothing happens until you write the code that calls it. How well it works on your data is yours to measure.
Pick Presidio if you write code and need to scrub a dataset or a set of transcripts before sending it to an AI service. Everyone else should pick an app.
The provider's own controls
Cost: The settings are free. Business plans are priced by OpenAI. Platform: any.
This is not redaction, and it is the first thing to fix. OpenAI's help pages describe three things worth knowing. Under Settings, Data Controls, "Improve the model for everyone" decides whether chats on a personal account can be used to train its models.
Temporary Chat is documented as not appearing in your history, not creating or updating memories and not being used to improve models while it stays temporary. However, OpenAI may keep a copy for up to 30 days for safety. For ChatGPT Business, Enterprise and Edu and for the API platform, OpenAI says business data is not used for training by default unless the customer opts in.
For regulated health data there is a further contract, the Business Associate Agreement. The addendum OpenAI publishes names parts of its API as eligible, and which ChatGPT plans qualify is a question for OpenAI and your counsel. Claude, Gemini and the others have their own settings.
What it's good at: it covers everything you type, costs nothing at the free end and settles the training question. What it lacks: your text still reaches the provider's servers and is kept for some period.
A setting is a promise, not a physical barrier. In 2025 a US court order in a newspaper lawsuit reportedly required OpenAI to preserve chats that users had deleted. The period reportedly ended in September 2025, with some of what was preserved still held. A business plan does not change what you are allowed to share. Your duties to a client, patient or student are the same either way.
Fix this layer first. For a team, a business plan with a signed agreement matters more than any redaction tool.
A local model
Cost: Free to download. Platform: macOS, Windows and Linux for both Ollama and LM Studio.
Ollama and LM Studio run open models on your own computer. LM Studio's documentation says it can operate entirely offline once a model is downloaded. Its privacy policy says messages, chat histories and documents are not transmitted from your system when you run models locally. Update checks and model downloads do use the network.
Ollama has a catch: it also offers cloud models, whose names end in :cloud, that run on Ollama's servers. Its documentation says prompts to those are processed there, and that cloud features can be turned off. Choose a local model if the point is that nothing leaves the machine.
What it's good at: it is the only option here where redaction is not needed, because there is no recipient. You can paste the whole note. What it lacks: setup time, memory and capability. Open models that fit on a laptop are generally less capable than the largest hosted ones, which matters for hard reasoning and less for summarizing or drafting. That is my judgment, not a benchmark.
Pick a local model if the material must stay on your machine and a smaller model is good enough for the job.
Find and replace by hand
Cost: Free. Platform: any text editor.
Copy the text into a plain-text editor. Replace each name with "Person A", "Person B" and each date with "Date 1", and keep a small key in a separate note. Ask your question, then reverse the replacements in the answer.
It works with any AI service, no vendor sees anything, and it forces you to read every line, which a tool does not. The cost is reliability on long text: a misspelled surname, a nickname or an employer slips through. Use it for short, occasional documents, or as the read-through after a detector's first pass.
Eight more tools
Each is real and useful to someone. None is on the shortlist because one of the seven above covers the same ground for most people, or the tool addresses a different problem.
Caviard
Cost: Free for up to 10 protected values per document. The lifetime plan is $9, which the page lists as an early discount from $15.30. Platform: Chrome extension for ChatGPT and DeepSeek.
It masks details in the browser before a prompt is sent, and Alt+R toggles original and redacted text. Its page describes detection as pattern-based, which suits emails and card numbers and tends to be weaker for names. I have not tested it on names. Pick Caviard if you use only ChatGPT or DeepSeek in Chrome and want a low-priced lifetime plan.
ClipScrub
Cost: $29 once, 14-day free trial. Platform: macOS 15 or later. The AI-enhanced detection needs Apple silicon.
ClipScrub is the closest Mac peer on price. It redacts PDFs, Word files, screenshots, screen recordings, JSON, CSV, XML and clipboard text, on the Mac and with no internet needed, according to its page. A keyboard shortcut starts it. It does not watch your typing, and it says it is not a compliance service and does not guarantee complete de-identification. Pick ClipScrub over Outbound Veil if your material is files, screenshots or recordings rather than text typed into boxes.
Clipmask
Cost: Free to download, with no subscription and no time limit. Its page also mentions a 500-copy trial limit, after which you may decide to continue, and gives no price for that. Platform: Apple silicon Macs only, macOS 14 or later.
Clipmask is the clipboard-only sibling of Predact, made by QCG. You copy text as usual, and a shortcut (Option+Space) offers a redacted version to paste. Its page says detection runs on the Mac, and that the only server contact is a check that the app is a licensed current version. Pick Clipmask if you want a free Mac tool that works on whatever you copy.
RedactDesk
Cost: Free, open source (MIT). Platform: macOS 14 or later.
RedactDesk is a Mac app for redacting PDFs, scanned ones included, with OpenAI's open-source privacy-filter model running locally. It deletes the text from the PDF rather than covering it. It handles PDFs only. Pick RedactDesk if your whole job is PDFs.
BlurData
Cost: $79 a year or $149 once for one person. Team plans are listed. No free tier is shown. Platform: macOS 13 or later, with an iOS version.
It finds and blacks out personal details in JPG and PNG screenshots and in PDFs, on the device, in batches. Pick BlurData if your leaks are screenshots.
MetaVeils
Cost: A free edition, and a Pro Lifetime License at $39.99 one time. Pro adds image OCR, deep metadata extraction and metadata stripping. Its page lists no limits for the free edition. Platform: macOS, as a .dmg download. The minimum macOS version and Apple silicon or Intel support are not stated.
MetaVeils is a Mac app from PholidLabs (version 1.1.3) for cleaning documents before you paste them into an AI chatbot. Its page lists TXT, PDF, DOCX and images. It describes finding names, email addresses, phone numbers, addresses, card and bank numbers, Social Security numbers and GPS coordinates. It also describes removing metadata such as EXIF, GPS and author fields.
The page calls it 100% offline, with scanning done on your Mac. I read that page on October 9, 2026 and have not installed the app, so that is the vendor's claim and I have not tested it. The page does not say whether it checks clipboard text or what you type. Pick MetaVeils if your material is documents and file metadata rather than text typed into boxes.
Nightfall
Cost: By quote, per user per year, in two editions. Directories quote conflicting figures, so I give none. Platform: browser, endpoint, SaaS and AI-app coverage, run by an administrator.
Nightfall is data loss prevention software for organizations. A security team sets policies, and the software blocks or flags sensitive data going into AI apps, email and other services. Pick Nightfall if you have to enforce a policy across many employees.
Microsoft Purview
Cost: Depends on your Microsoft licensing, which I did not confirm. Platform: managed devices, Edge and other browsers.
Microsoft's documentation describes Endpoint DLP (data loss prevention) policies that block pasting and uploading sensitive information to AI app websites on managed devices. It also describes Edge policies that inspect prompts typed or pasted into consumer AI apps including ChatGPT, Copilot, DeepSeek and Gemini. Pick Purview if your organization already runs Microsoft 365 security and IT can turn it on.
The seven narrower options are mostly young, with a handful of installs or stars as of this writing. ChatGPT Privacy Shield is a free Chrome extension for ChatGPT only, with Pro ($9 a month) and Team ($49 a month) plans listed as a waitlist. Rescriber is a free, open-source ChatGPT extension from Northeastern University researchers. PII Scrubber (ChatGPT, free for 5 uses a day, Pro $7 a month) and Paste Redactor (four browsers, free for 100 redactions a month, then $1 a month) are small extensions with paid tiers.
PIIGuardAI is an open-source Mac menu-bar proxy that blocks AI requests containing patterns such as emails and card numbers but uses regular expressions only. Maskbase is an open-source Mac app for documents. Veil (helloveil.com, $8 a month) is a different company's meeting recorder with a privacy engine. Its changelog lists nothing that checks what you type in other apps.
Three more were added on October 14, 2026, each as listed that day and not tested by us. Data Guard and Prompt Guard are Firefox add-ons for AI chat sites, which matters if you use Firefox, since several tools above are Chrome extensions. Data Guard warns without blocking, and Prompt Guard masks matches. Neither listing states a price. Whiteout is a Mac menu-bar app for screenshots, and its only listing I could read is on hunted.space. Mac screenshot apps cover images and Outbound Veil covers typed text, so they do not replace each other.
Two more Mac apps were added on October 21, 2026, each as listed that day and not tested by us. Redaktr checks screenshots on the Mac before you share them, with a free trial and then $39 once, as listed on Product Hunt. MaskPaste masks names, companies and addresses before you paste into a chatbot and swaps the originals back into the reply, as listed on hunted.space; it does not state a price there. Outbound Veil checks typed text and does not swap originals back.
One more was added on November 4, 2026, as listed that day and not tested by us. SinapticAI is a free browser extension from the Mac App Store that, per its listing, warns before you paste personal or financial details into a website. Its last update was January 7, 2026. It watches pastes in the browser; Outbound Veil checks what you type in any app.
What redaction cannot do
Some vendors above say this plainly, and everyone should assume it. Rampart, the model behind Outbound Veil, calls itself harm reduction, not perfect protection, and Presidio says there is no guarantee it finds everything.
- It lowers the chance of a mistake. It is not compliance. No tool here makes a workflow compliant with a regulation or a contract.
- It misses things. Examples include nicknames, initials and a case number in a format the detector has not seen. Read the text yourself before you send it.
- It flags the wrong things too. A detector cannot tell your client from a famous person, so public figures get flagged. You ignore those.
- Non-Latin scripts are a gap. Rampart's documentation reports about 98% recall (the share of real names the model finds) for names in seven Latin-script languages. However, it reports about 14% across non-Latin scripts in aggregate, from 2% for Russian to 46% for Japanese. Any tool built on it, Outbound Veil included, inherits that. Other tools use other models. Test each on your languages.
- Context can identify a person with no name in sight. Examples include a rare diagnosis, a job title at a small employer, a town and a date. Detectors are built around identifiers such as names and numbers, and Rampart's documentation lists indirect identifiers as unaddressed.
- Text tools do not touch files. Attachments, screenshots and metadata need a file tool, or should not be uploaded. Nor is redaction retroactive.
HIPAA, FERPA and client confidentiality
This is not legal advice, and the answer depends on who you are and what you hold.
HIPAA (the US Health Insurance Portability and Accountability Act). The rule that defines de-identified health information (45 CFR 164.514) offers two routes. The first is an expert's documented determination that the risk of identifying a person is very small. The second is removal of a listed set of about 18 identifier types, with no actual knowledge that what is left could identify someone. The listed types include names, geographic detail, dates tied to a person, contact numbers and record and account numbers.
A detector that catches names and numbers is neither route, so text it has cleaned can still be protected health information. Whether you may send data to an AI service at all usually turns on a Business Associate Agreement and your organization's policy, so ask your compliance officer.
FERPA (the Family Educational Rights and Privacy Act). The law gives parents, and students once they are 18 or in postsecondary school, rights over personally identifiable information in education records. If you grade, advise or write recommendations, check your institution's AI policy before pasting anything that names a student. Removing the name is a habit, not a rule.
Client and participant confidentiality. Lawyers, therapists, consultants and researchers have duties in codes of conduct and contracts. The American Bar Association's Formal Opinion 512 (July 2024) is reported to say that a lawyer's confidentiality duty applies to generative AI tools. It is also reported to say that tools which train on inputs call for the client's informed consent. In California, SB 574 adds section 6068.1 to the Business and Professions Code from January 1, 2027: a lawyer must not enter confidential, personal identifying or other nonpublic information into a generative AI system unless access to it is restricted to the lawyer and people bound to keep it confidential (the chaptered text, read October 21, 2026; what it lists as personal identifying information). This is not legal advice.
For research, your IRB (institutional review board) protocol and data management plan probably already say who may see participant data, and an AI service is a new recipient. For manuscripts, see what publishers allow when you run your own manuscript through AI.
Decision matrix
- Mac user typing or pasting client or patient text into chat boxes? Outbound Veil, $29.99.
- Documents, scans or screenshots, or on Windows, or want it free? Predact. On a Mac, ClipScrub too.
- Several AI sites in a browser, values restored in the reply? PrivacyScrubber.
- Only ChatGPT in Chrome? Caviard, Privacy Shield or Rescriber.
- Only PDFs? RedactDesk. Mostly screenshots? BlurData or ClipScrub. Whatever you copy? Clipmask.
- A developer scrubbing a dataset or transcripts? Presidio.
- A team with a policy to enforce? A business plan first, then Nightfall or Purview.
- The data must not leave the machine? A local model.
- Occasional short text, no new software? Find and replace.
A routine that holds up
- Write down what must never go in, then fix the account. Turn off the model-improvement setting or use a business plan, and use Temporary Chat for one-off questions.
- Put a detector where you type. A menu-bar app or an extension, or a scrubber page you open every time.
- Test it with fake data first. Paste a made-up record with a name, a number, an address and a date, and see what it catches. This applies to every tool here, including mine, and takes two minutes.
- Read what you are about to send. Look for what a detector would not know: a nickname, an employer.
- Ask with placeholders and restore by hand. Keep any key in a separate local note. For the most sensitive material, use a local model or skip AI.
Frequently asked
What is the best way to remove personal information before using ChatGPT?
Use layers. Set the account so chats are not used for training, or use a business plan. Put a detector where you type. Read what you are about to send. For material you cannot send at all, use a local model.
Does ChatGPT Temporary Chat keep my data private?
Only partly. OpenAI says a temporary chat stays out of your history, creates no memories and is not used to improve its models while it remains temporary. However, it may keep a copy for up to 30 days for safety. The text still reaches OpenAI's servers.
Is a redaction tool enough for HIPAA?
No. HIPAA defines de-identified data by an expert determination or by removing a listed set of identifiers with no actual knowledge that the rest could identify the person. A tool that catches names and numbers is neither. This is not legal advice. Ask your compliance officer or counsel.
Do these tools send my text to a server?
Most say detection runs on your device or in your browser. That is the vendor's claim, not something I tested. Enterprise DLP and the provider's own plans involve servers by design. A local model keeps text on your machine.
Is there a free tool to redact personal information for AI chats?
Yes. Predact (Mac and Windows) is free, and Clipmask (Apple silicon Macs) is free to download, with a 500-copy trial limit mentioned on its page. RedactDesk is a free, open-source Mac app for PDFs. Rescriber is a free, open-source Chrome extension for ChatGPT. PrivacyScrubber and Caviard have free tiers with limits.
Is there a Mac app that checks for personal information as I type?
Outbound Veil, which Purplelink makes, checks the text field you are typing in, in the apps it can read. It costs $29.99 once with a 7-day trial. Other Mac tools work on what you copy or drop in: Clipmask, ClipScrub and Predact. I found no other that watches the typing itself, but I may have missed one. Outbound Veil is Mac only, recognizes names in Latin script only, and will miss some things.
Should I run a local model instead of redacting?
If the material must not leave your machine, yes. Ollama and LM Studio run open models on your computer. The cost is setup time and usually a less capable model. Ollama also offers cloud models that run on its servers, so choose a local one.
Related guides
- Before you paste client notes into ChatGPT or Claude: a checklist for Mac users: a five-step checklist.
- Can you use AI to check your own manuscript? What publishers allow.
- Do researchers need a VPN?: what a VPN hides and what it does not, for sensitive research data.
- Best Mac apps for transcription, text to speech and OCR: if the text you will paste into a chat comes from a recording or a scan.
- Purplelink's privacy page, the Mac Suite and all guides.