get ancesra

tell us where to send it. the download link arrives by email in a minute.

free, forever. we keep your name and email privately, to know who uses ancesra and to contact you about it. never sold, never shown in public. the app itself has no account.

docs

everything you need to get ancesra running and get the most out of it. short version: download, open, wait for the one-time setup, then unplug the wifi.

Install

  1. Get the installer. Click download and enter your name and email: the download link arrives by email. The newest version is 2.6.0 (121 MB, Windows 10 or 11).
  2. Run it. Windows may warn that the app is from an unknown publisher, because the installer isn't code-signed yet. Choose More info, then Run anyway.
  3. Open Ancesra with the internet on, once. The first start installs Ollama (the free program that runs the AI) and downloads the AI model, about 2.5 GB. The loading screen shows each step.
  4. That's it. From now on Ancesra works with the internet off.
download for windows

What your PC needs

  • Windows 10 or 11, 64-bit. A Mac version isn't available yet.
  • About 6 GB of free disk space for the app, Ollama and the AI models.
  • 8 GB of memory or more is comfortable. A graphics card (GPU) makes answers faster but isn't required.
  • Internet once, for the first setup. After that, never required.

Chat, coding, debug

Pick a mode at the top of the chat. Each one shapes the answer differently.

Chat

General questions. A direct answer first, then a few key points. It knows events up to 2026 from a built-in Wikipedia pack and shows its sources.

Coding

Name the function and the language. You get the approach, the full code, then its time and space complexity.

Debug

Paste broken code. One sentence on what's wrong, then the fixed code, keeping the same function name and parameters.

Settings lets you choose how detailed answers are (concise, balanced or detailed) and which mode Ancesra starts in.

My files

Add your own documents and Ancesra answers from them, naming the file and the page.

  • Add files from My files in the sidebar, or drop them anywhere on the window.
  • Supported: PDF, Word (.docx), .txt, .md, .csv and .json, up to 50 MB each.
  • Not yet: scanned PDFs (pictures of pages), because they have no text to read.
  • Private: files are read and indexed on your computer. Removing a file deletes its index; the original is never changed.

Try: how many sick days do I get? after adding a company handbook, or when is my midterm? with a course plan.

Memory and profile

  • Say remember that my exam is on Friday. Ancesra asks before saving anything about you.
  • Your Profile shows what it remembers, and you can delete any of it.
  • The profile also shows your stats: questions asked, active days, streak and the topics you ask about.

What goes online

Your chats, files, profile and memory never leave your computer. Ancesra only uses the internet for:

  • The first setup: downloading Ollama and the AI model.
  • Checking for updates, when you're online: one request to GitHub asking for the newest version. No chats or files are sent.

The app has no account and no analytics, and no server of ours is involved in running it.

Downloading from this website is different. It asks for your name and email, which are kept privately and never sold or published; the email is used to send you the download link and to contact you about Ancesra. Separately, each download is counted in a public log (download.md) with the time and your approximate place (city and country, worked out from your IP address). The IP address itself is never stored.

Updates

When a new version is out and you're online, Ancesra offers it in a dialog. Choose Download, then Restart and update: it installs itself and opens again, with your chats kept. You can also use Check for updates at the bottom of the sidebar, or skip a version.

Troubleshooting

The loading screen says Ollama didn't start
Restart your computer and open Ancesra again. If it still fails, reinstall Ollama from ollama.com.
It says it needs to download the AI model
The first setup needs the internet once (about 2.5 GB). Connect, reopen Ancesra, and let it finish.
Another program uses port 3000
Nothing to do: Ancesra picks a free port by itself.
Answers are slow
The first answer after starting loads the model into memory. Closing heavy apps helps on PCs with 8 GB of memory.

FAQ

Is it really free?
Yes. No account, no subscription.
Which AI model does it use?
Microsoft's phi4-mini, an open model (MIT licence), running on your computer through Ollama.
Is there a Mac or Linux version?
Not yet. Windows only for now.
Can it browse the internet?
No. It answers from its own knowledge, the built-in knowledge pack and the files you give it.