Skip to main content
Before your question and the matching passages from your documents go to cloud AI, Elephas finds many common personal details on your computer and replaces them with placeholders. When the answer comes back, Elephas puts the real words back in, on your computer.
Protection applies to Elephas AI and your own API key. Local AI sends nothing to the cloud, so there is nothing to hide from it.

​
Before you start

  • The private-information models must be downloaded. They download during first-run setup and appear in Settings › Downloads as Private-information models. See First-run downloads.
  • In a Team workspace, the Team’s AI policy decides whether redaction runs, not your personal setting. See Team AI policy.

​
Steps

​
Turn protection on or off

1

Open Settings › AI

Open Settings and choose AI.
2

Find the switch

Find Protect personal information when using cloud AI. It is on by default.
3

Leave it on, or turn it off

If you turn it off, your question and sources are sent to cloud AI exactly as they appear in your documents.
The steps are the same on Mac and Windows.

​
Always hide specific names

Automatic detection can miss things. To make sure a client name, project code name or other term is always hidden, add it to the Brain’s Protected Names: “Names and terms listed here are always replaced with placeholders before this Brain’s content is sent to cloud AI.” See Protected Names.

​
Check what was hidden

Under each answer protected this way, a shield shows how many details were hidden, for example 3 details hidden. Click it to see the list. See Review hidden details.

​
How it works

​
What Elephas hides

Elephas looks for these kinds of personal details. The names below are the ones you see in the shield list. Plus anything you add to a Brain’s Protected Names. Job titles and honorifics (such as “Dr.” or “CEO”) are deliberately not hidden, so the AI can still understand who is who.

​
How it runs on your computer

Detection happens on your computer, before anything is sent. Elephas uses three methods together:
  1. Protected Names you add to a Brain. Matched as whole words, ignoring case and accents.
  2. Patterns for structured details such as email addresses, card numbers, IBANs and phone numbers. Card and Aadhaar numbers are also checked with their checksums.
  3. A small on-device model that recognises names, organisations, places and similar details in running text. It runs locally and is downloaded once during setup.
Elephas checks your question, the chat history it sends, and every passage it sends. Each detail is replaced with a placeholder such as [PERSON-3fa2c1]. The same name always gets the same placeholder within a Brain, so the AI can still tell that two passages are about the same person. It does not see who that person is. The link between placeholders and real words is kept in memory only. It is never saved, synced or written to a log.

​
If the models are not ready, nothing is sent

If protection is on but the private-information models have not finished downloading, Elephas does not fall back to weaker protection. It stops and shows: “Elephas can’t protect personal information until it finishes downloading. Nothing was sent. Open Settings to finish setting up, or turn the protection off.” If protection fails for any other reason, Elephas shows: “Personal information couldn’t be protected for this answer, so nothing was sent.”

​
Limits you should know

  • It hides many common personal details, not all of them. The model is a statistical detector. In internal testing on English text it found about 94% of the personal details in the test set. That means some will be missed. Unusual names, nicknames, and details that only identify someone in context (a job title plus a small town, for example) can slip through.
  • It sometimes hides things that are not personal. That is the cautious side of the tradeoff. The real words are restored in your answer.
  • English is the supported language. The model is built for English text. Patterns such as email addresses, card numbers and IBANs are detected in any language, but names, organisations and places in other languages may not be recognised.
  • Street addresses and usernames are not reliably caught. Add the ones that matter to Protected Names.
  • The topic of your question is still sent. Placeholders hide who and which, not what. The AI still reads the rest of the passage.
For material where any miss is unacceptable, use Local AI. In Local AI mode, your question and your documents are not sent to any AI provider.

​
Troubleshooting

The private-information models are missing or still downloading. Open Settings › Downloads, check Private-information models, and finish the download. Then ask again. See First-run download failed.
One of the documents in scope is not in English, so Elephas could not protect it and sent nothing. Ask about English documents only, switch to Local AI, or turn protection off if you accept sending that text as it is.
Add it to the Brain’s Protected Names. From then on it is always replaced before that Brain’s content is sent to cloud AI.

​
FAQ

No. It hides many common personal details, listed above, and it will miss some. You can see exactly what was hidden for each answer with the shield, and add anything that matters to Protected Names.
No. Detection runs on your computer with a model stored on your computer. Only the text with placeholders is sent to the AI you chose.
Yes. Protection applies to both Elephas AI and your own API key, with every supported provider and custom endpoints.
Elephas puts the real words back on your computer as the answer arrives. The AI only ever wrote the placeholders.
In a Team workspace, the Team Owner sets the AI policy. It can disable cloud AI, allow it, or allow it only with required redaction. Only the Team Owner can change it. See Team AI policy.

What data is sent to AI

Review hidden details

Protected Names

Run Local AI