AI Prompt Privacy Checker
Automatic prompt privacy scanning
You want to detect common sensitive data, review or restore replacements, and manually label anything missed before sending text to AI.
AI privacy guide
AI privacy starts before you paste, upload, or attach
A practical guide to what prompts and files can expose to AI tools, and how to inspect, clean, or redact that data locally before sharing.
Last reviewed:
If a prompt, screenshot, image, PDF, or document contains private data when you send it to an AI assistant, assume that data can be processed by that provider according to its current settings and policies.
The practical fix is not panic. It is review. Check what the content reveals, remove what should not travel, then share only the cleaned copy or prompt with detected sensitive data replaced by labels.
Freshmii focuses on that review step. The tools run in your browser, so your files and text are never uploaded to a Freshmii server.
Visible data is what you can see on the page or screen, such as names, emails, account numbers, addresses, signatures, screenshots, and PDF body text.
Hidden data is carried inside the file, such as GPS coordinates, EXIF fields, author names, timestamps, document properties, comments, and revision signals.
A complete review checks both layers. Redacting visible content does not automatically remove hidden metadata, and removing metadata does not hide sensitive text that is still visible.
Before sending a prompt, look for client names, credentials, IDs, project codenames, internal URLs, personal data, and anything that gives the AI more than it needs.
Before uploading an image or screenshot, check visible details first, then check whether the file carries location, device, author, or editing metadata.
Before attaching a PDF or document, review visible body text, then check document properties, comments, author fields, timestamps, and other hidden metadata.
AI providers offer different privacy controls depending on whether you use a consumer account, business workspace, enterprise plan, or API. Those settings can change over time.
Provider settings are still downstream controls. The strongest step you control directly is reducing what you send in the first place.
This page links to provider privacy documentation for reference, but you should always verify the current policy for the account and product you actually use.
Use this matrix to choose the first cleanup step before content reaches ChatGPT, Claude, Gemini, Copilot, or another AI assistant. It separates visible data from hidden file metadata, then maps each risk to a local Freshmii workflow and a verification step.
| Content type | Visible risk | Hidden risk | Best Freshmii workflow | Verification step |
|---|---|---|---|---|
| Prompt or pasted text | Names, client context, credentials, internal IDs, private URLs, support notes, legal or HR text. | Hidden Unicode, copied formatting, and private context mixed into logs or document extracts. | Clean prompt before AI: automatically detect common sensitive data, review replacements, and manually label anything missed before sending the prompt. | Copy the cleaned text into a separate note and search for the original names, emails, API keys, IDs, and private URLs. |
| Screenshot | Emails, account numbers, customer names, API keys, internal URLs, balances, or private UI state. | Original image metadata can remain unless the output is re-encoded or inspected. | Redact screenshot before uploading to ChatGPT or another AI assistant: black out visible sensitive regions, then review the exported image. | Open the exported image outside Freshmii, zoom in, and confirm covered pixels cannot be read by visual AI analysis. |
| Photo or image file | Faces, addresses, documents, license plates, labels, screens, or private places. | GPS coordinates, capture time, device model, author/copyright fields, editing history, AI generation metadata. | Remove metadata before uploading to AI: inspect EXIF, GPS, device, timestamp, and author fields, then strip sensitive fields from a cleaned copy. | Run the cleaned file through Metadata Inspector again and confirm location, device, timestamp, and author fields are gone. |
| Body text, names, signatures, account numbers, case details, pricing, medical or financial data. | Author, creator, producer, creation/modification dates, form fields, comments, attachments, or document properties. | Use the browser-local PDF redactor for visible content, then inspect or remove document metadata from the exported copy. | Open the redacted PDF, search for removed values, try selecting text under black boxes, and inspect metadata separately. | |
| DOCX or document package | Document body text, headers, footers, notes, client details, tracked-change content, or comments. | Author, company, manager, template, timestamps, custom properties, comments, relationship targets, media names. | Check file metadata before AI upload: inspect author fields, comments, tracked-change signals, timestamps, and custom properties before sharing. | Create a cleaned copy, reopen it, and inspect metadata again before sending the document or excerpt to an AI tool. |
This printable checklist turns the risk matrix into a before-upload workflow for prompts, screenshots, photos, PDFs, DOCX files, support notes, and debugging logs. Use it to reduce what you share, clean prompt before AI, remove metadata before uploading to AI, and verify the cleaned copy before it reaches an AI provider.
Download checklist PDF Use the risk matrix
| Review stage | What to check | Freshmii workflow |
|---|---|---|
| Reduce the upload | Start by sending the smallest useful prompt, excerpt, screenshot, or file.
| Use the AI Prompt Privacy Checker for pasted prompts, support notes, logs, and document excerpts before asking an AI assistant for help. |
| Scan visible text | Find sensitive text before it reaches ChatGPT, Claude, Gemini, Copilot, or another AI system.
| Use AI Prompt Privacy Checker for detection and AI Text Cleaner when copied text also contains markdown, hidden Unicode, or messy formatting. |
| Redact screenshots and images | Cover visible sensitive regions and avoid sending hidden image metadata by mistake.
| Use Screenshot Redactor for visible pixels, then Metadata Inspector or Metadata Remover when image metadata also matters. |
| Check PDFs and documents | Treat visible redaction and hidden document metadata as separate checks.
| Use PDF Redactor for visible PDF content, then Metadata Inspector and Metadata Remover for document properties and hidden fields. |
| Verify before upload | Confirm the cleaned copy is the only version that will be sent to the AI provider.
| Keep the original and cleaned copy clearly separated, then upload only the verified cleaned version. |
The PDF includes provider references for checking current AI data controls: OpenAI ChatGPT privacy practices, Anthropic Claude model-training data use, Google Gemini Apps Privacy Hub.
AI Text Cleaner
Cleaning AI output before reuse
The issue is markdown, hidden characters, spacing, or copy-paste formatting.
Metadata Inspector
Finding hidden file data
You need to know what metadata a file exposes before deciding what to remove.
Metadata Remover
Removing hidden file metadata
You want a cleaned copy without GPS, author, timestamp, PDF, or DOCX fields.
Screenshot Redactor
Blackout redaction in images
The sensitive data is visible in a screenshot or flat image.
PDF Redactor
Visible PDF redaction
Names, account numbers, signatures, or other visible PDF content should be covered before sharing.
Freshmii articles stay tied to the same theme as the tools: what private data is exposed, what to clean before AI sees it, and which browser-local workflow fits the job.
AI privacy basics
AI Data Retention Questions Before Sharing
A practical checklist for reviewing AI data retention before sharing prompts or files. Learn which provider settings, account types, storage periods, model-training controls, and audit realities matter before private content leaves your browser.
AI privacy basics
AI Privacy Checklist Before Using ChatGPT, Claude, or Gemini
A before-AI privacy workflow for reducing exposure before content reaches an AI provider. Decide what the model actually needs, use automatic detection plus review controls to replace sensitive strings with placeholders, check visible and hidden file data, verify the cleaned copy, and then review the privacy controls for the account you are using.
AI privacy basics
Browser-Local vs Cloud AI Privacy Tools
A comparison of browser-local and cloud AI privacy tools. Learn when local cleanup is safer, when cloud services may still be useful, and how to decide which workflow fits prompts, files, screenshots, PDFs, and metadata.
AI privacy basics
Browser-Local vs Extension vs Cloud Privacy Tools
A decision guide for choosing between browser-local web tools, browser extensions, and cloud privacy tools before using AI. Compare data flow, permissions, repeat-use convenience, file support, audit needs, extension host access, and when each approach fits prompts, screenshots, PDFs, metadata, and team workflows.
AI privacy basics
Verify Browser-Local Processing With the Network Tab
A screenshot-backed verification guide for checking Freshmii's browser-local AI privacy tools. Learn what network activity is expected, what would be suspicious, and how to confirm there is no upload request after scanning, redaction, or metadata cleanup.
Prompt privacy
Anonymize Customer Support Prompts Before AI
A support-team guide to anonymizing customer prompts before AI. Learn how to sanitize customer email before AI, replace identifiers with useful placeholders, preserve support context, avoid leaking screenshots, and review AI output before sending it back.
Prompt privacy
Remove Secrets From Code Before Asking AI
A developer-focused guide to cleaning code, logs, config, .env files, and stack traces before AI assistance. Use a practical checklist, safe placeholders, and before-and-after log examples to remove API keys before asking AI and sanitize logs before AI debugging.
Prompt privacy
What Not to Paste Into AI Prompts
A practical guide to prompt privacy. Learn which categories of text should be treated as no-paste by default, how automatic detection plus review controls can replace sensitive strings with useful placeholders, and when a sanitized prompt is safer than a full copy of the original note, log, email, or document excerpt.
File upload privacy
AI File Upload Privacy Checklist for Teams
A team-oriented checklist for AI file uploads. Learn how to classify files, inspect visible and hidden data, decide what to redact, verify cleaned copies, and document the minimum review step before sharing with AI.
File upload privacy
Check Files Before Uploading Them to AI
A file-upload privacy guide for AI workflows. Learn how to separate visible content from hidden metadata, how to triage images, screenshots, PDFs, and DOCX files, and which local Freshmii tools to use before a file reaches an AI provider.
File upload privacy
Convert Files to Excerpts Before Uploading to AI
A guide to replacing full file uploads with sanitized excerpts. Learn when to paste text instead of uploading a document, how to preserve useful context, and when a cleaned file is still the better workflow.
File upload privacy
Supported File Types and Privacy Fields
A source-backed reference for Freshmii supported file types and privacy fields. Compare which formats can expose GPS, author names, timestamps, document properties, visible pixels, AI workflow metadata, comments, tracked changes, credentials, and other sensitive data before AI upload.
Image metadata privacy
Remove GPS From Photos Before Uploading to AI
A focused guide to GPS privacy in photos before AI upload. Learn how EXIF location fields work, why screenshots and edits are not enough to assume safety, and how to inspect, remove, and verify GPS metadata locally.
Image metadata privacy
Remove Photo Metadata Before Uploading Images to AI
A practical guide to image metadata privacy. Learn what EXIF, GPS, IPTC, and XMP fields can reveal, when metadata removal is useful, why visible redaction is a separate step, and how to inspect or clean supported image metadata locally.
Image metadata privacy
Strip AI Image Prompts and Workflow Metadata
A guide to AI image metadata before upload or reuse. Learn what prompts, seeds, workflow data, C2PA-style provenance, and editor metadata can expose, and how to inspect and remove sensitive fields before sharing with AI.
Image metadata privacy
What Phone Photo Metadata Exposes Before AI
A source-backed phone photo metadata example for AI privacy reviews. See realistic EXIF, GPS, IPTC, XMP, author, timestamp, camera, software, and technical fields, what each field can reveal, and how to verify a cleaned copy before uploading to ChatGPT, Claude, Gemini, Copilot, or another AI assistant.
Screenshot redaction
Blur vs Blackout Screenshot Redaction
A screenshot redaction guide comparing blur, pixelation, cropping, and solid blackout before AI upload. Learn why reversible or weak visual masking is risky, when cropping is better, and how to verify the exported image.
Screenshot redaction
Redact Browser Screenshots Before AI Uploads
A browser-screenshot privacy guide for AI workflows. Learn what browser chrome and web-app UI can reveal, how to crop or black out sensitive areas, and how to verify the exported screenshot before upload.
Screenshot redaction
Redact Screenshots Before Asking AI to Analyze Them
A screenshot privacy guide for AI workflows. Learn what screenshots commonly reveal, how to redact visible details before upload, when cropping is not enough, and how to verify an exported image before asking an AI assistant to analyze it.
PDF redaction
Redact PDFs Before Uploading Them to AI
A PDF privacy guide for AI uploads. Learn why visible redaction and metadata cleanup are separate jobs, how to verify a redacted PDF, and when to use local PDF redaction before sharing statements, contracts, forms, reports, or case documents.
PDF redaction
Remove PDF Metadata After Redaction
A guide to PDF metadata cleanup after visible redaction. Learn what PDF document properties can reveal, why redaction and metadata removal are separate, and how to inspect, strip, and verify hidden fields before AI upload.
PDF redaction
Verify PDF Redaction Before Uploading to AI
A PDF redaction verification guide for AI workflows. Learn how to check visual redactions, searchable text, metadata, filenames, comments, and copied output before uploading a PDF to an AI assistant.
Document metadata privacy
Clean DOCX Metadata Before Sharing Documents with AI
A DOCX privacy guide for AI workflows. Learn what document metadata and hidden parts can reveal, how to inspect supported DOCX fields locally, why metadata cleanup does not rewrite visible text, and when to sanitize excerpts instead of uploading a full document.
Document metadata privacy
Remove Comments and Tracked Changes Before AI
A document privacy guide for comments and tracked changes before AI upload. Learn why final-looking DOCX files can still contain review history, how to inspect hidden parts, and when to export a clean excerpt instead.
Document metadata privacy
Sanitize Resumes and HR Documents Before AI
An HR privacy guide for using AI with resumes and employee documents. Learn what to remove, when to use placeholders, how to avoid hidden DOCX metadata, and how to review AI output before it becomes part of a people process.
AI output cleanup
Check AI Output for Private Data Echoes
A guide to reviewing AI-generated output for private data echoes. Learn why sanitizing the input is not enough, what strings to search for, how to handle placeholders, and how to clean formatting without skipping human review.
AI output cleanup
Clean AI Output Before Reusing It
A practical guide to cleaning AI-generated text before reuse. Remove markdown clutter, hidden characters, copied HTML, filler language, accidental private echoes, and placeholder mistakes while keeping a clear review boundary between formatting cleanup and factual approval.
AI output cleanup
Remove Hidden Unicode From AI Output
A guide to hidden Unicode in AI output. Learn what invisible characters can do, why they matter for documents, prompts, code, forms, and publishing, and how to clean text before reuse without changing its meaning.
Provider policies change, and the right setting depends on the product tier you use. These links point to provider documentation and are starting points, not a substitute for checking your own account controls.
OpenAI / ChatGPT
Data Controls let users manage whether ChatGPT conversations help improve models; OpenAI says business products such as ChatGPT Business, Enterprise, and API are excluded from training by default unless opted in.
Anthropic / Claude
Claude privacy settings and retention can differ between consumer products and commercial/API services. Check the current model-improvement setting for the account you use.
Google / Gemini
Gemini Apps Activity, temporary chat, feedback, and workspace products have different data-use behavior. Review the Privacy Hub for the current controls.
- What is AI privacy in practical terms?
- AI privacy means limiting what private data reaches an AI provider. Before you paste, upload, or attach content, inspect visible and hidden data, remove what is unnecessary, and share only the cleaned version.
- Is metadata as important as visible text?
- Yes. Visible text and hidden metadata are separate layers. A photo can look harmless but carry GPS coordinates; a PDF can have redacted-looking pages while still carrying author or creator fields.
- Do browser-local tools replace provider privacy settings?
- No. Browser-local cleanup reduces what you send. Provider settings control what happens after the data reaches that provider. Use both when the content matters.
- Does Freshmii upload my files for processing?
- Freshmii tools run entirely in your browser. Your files and text are never sent to a Freshmii server. You can verify this by watching the Network tab while using a tool — nothing leaves your device.
Ready to review content before AI sees it?