Before you start
- Local AI needs enough memory and disk space. Models range from about 1.2 GB to 5.3 GB. See Choose a local model.
- Downloads need a connection once. The first-run models and the Local AI model must be downloaded before you go offline. See First-run downloads.
- Personal workspace. In a Team workspace, the Team Owner’s AI policy decides which modes are allowed. See Team AI policy.
Steps
1
Set up Local AI
Open Settings › AI. Under How Elephas answers, choose Local (on this computer). Pick a model or keep Choose for me, then confirm the download. Wait until the status reads Local AI ready. Full steps: Run Local AI.
2
Check that every download is finished
Open Settings › Downloads. When everything is present, it reads “Everything is downloaded. Elephas works offline.”
3
Leave Personal Sync off
Open Settings › Sync. If Sync is on, see Turn off Sync. Sync is off unless you turned it on.
4
Turn off usage reports (optional)
Open Settings › Privacy and turn off Send anonymous usage and error reports. Reports never include your documents or questions, but this stops them too. See Usage reports.
5
Check an answer
Ask a question. The answer’s badge reads Answered by: This computer.
How it works
What stays on your computer
Indexing happens on your computer. Always. In every mode, reading your files, OCR, the search index and finding passages run locally. In Local AI mode, answering does too. The model runs on your computer, through a local engine that only accepts connections from your own computer. Your question, the passages it reads, your chat history and the answer are not sent to any AI provider. In the app: “Answers are generated on this computer. Nothing is sent anywhere.”What still goes online, and why
With Sync off, nothing is sent to the sync service.
Working with Wi-Fi off
Turn off your Wi-Fi after Elephas has confirmed your plan, and Local AI still answers from your documents. Elephas keeps a confirmed plan check in memory for up to an hour. After that hour, or after you restart Elephas, it needs to reach the Elephas service again before the next Local AI answer. If it can’t, it shows: “Elephas could not confirm your plan right now. Check your connection and try again.” Your documents, Brains and chats stay available to browse.Troubleshooting
Elephas could not confirm your plan right now
Elephas could not confirm your plan right now
Elephas needs to reach the Elephas service about once an hour, and after each restart, to confirm your plan for Local AI. Reconnect briefly, then ask again.
Local AI isn't ready yet. Check Settings.
Local AI isn't ready yet. Check Settings.
The model has not finished downloading or setting up. Open Settings › AI and check the Local AI status. See Local AI troubleshooting.
Answers are slow
Answers are slow
Local models are slower than cloud AI, especially on computers with less memory. A smaller model answers faster. See Choose a local model.
