AI mistress with access to your files
Posted: Thu Aug 21, 2025 11:29 am
I'm posting this here almost because I need help setting up the AI. If you're good at this and want to help, PM me (my discord is below)
This is a local AI chat where you can give your files to the AI.
At the beginning, you yourself set the folder to which the AI will have access. Outside this folder, the AI knows nothing.
But there is a limit! The folder should not contain more than 100 files,
and also for each request the AI will know about the existence of only 10 random files in this folder!
AI has the character of the dominant woman
(almost like me)
The program works on Windows, Linux and macOS (I only tested it on Windows, but it should work on others too)
The program consumes about 5 GB of RAM during operation! If you don't have that much, don't run it!
What AI can do with files:
1. Create files
2. Modify files (don't rename)
3. Delete files
4. Delete folders
This chat can have a lot of bugs if you give the AI unexpected requests.
What you need to start a chat (!IMPORTANT!):
1. Java 21 or later (You can download it here: https://www.oracle.com/java/technologie ... ds/#java21)
2. Ollama and downloaded model llama3.1:8b (How to download it below)
3. Run AnutAI.jar by double clicking (you can't do it, then input this in the console: java -jar "Full Path To Jar File")
How to install Ollama:
1. Download the program here: https://ollama.com/
2. When you download this program, open the console and input this: ollama run llama3.1:8b
3. The model that weighs 5 GB will start downloading
4. When the model is downloaded then you can close the console
5. But for AI chat to work, Ollama needs to be running! You can see this in the taskbar
I also need help!
If you are good at AI prompts, then DM me (discord: anutays)
Also for those who have a very powerful computer:
You can find and download other models on the Ollama website that are much more powerful,
such as llama3.1:70b (requires about 40 GB of RAM), or other models.
To run a program with a different model, you need to specify it when you run the jar file.
For example: java -jar "Full Path To Jar File" llama3.1:70b
Download: https://drive.google.com/file/d/1yNOqHQ ... sp=sharing
Mirror: https://mega.nz/file/rrYHDD7A#AxSSb1bSZ ... fjJKS0HUUA
This is a local AI chat where you can give your files to the AI.
At the beginning, you yourself set the folder to which the AI will have access. Outside this folder, the AI knows nothing.
But there is a limit! The folder should not contain more than 100 files,
and also for each request the AI will know about the existence of only 10 random files in this folder!
AI has the character of the dominant woman
(almost like me)The program works on Windows, Linux and macOS (I only tested it on Windows, but it should work on others too)
The program consumes about 5 GB of RAM during operation! If you don't have that much, don't run it!
What AI can do with files:
1. Create files
2. Modify files (don't rename)
3. Delete files
4. Delete folders
This chat can have a lot of bugs if you give the AI unexpected requests.
What you need to start a chat (!IMPORTANT!):
1. Java 21 or later (You can download it here: https://www.oracle.com/java/technologie ... ds/#java21)
2. Ollama and downloaded model llama3.1:8b (How to download it below)
3. Run AnutAI.jar by double clicking (you can't do it, then input this in the console: java -jar "Full Path To Jar File")
How to install Ollama:
1. Download the program here: https://ollama.com/
2. When you download this program, open the console and input this: ollama run llama3.1:8b
3. The model that weighs 5 GB will start downloading
4. When the model is downloaded then you can close the console
5. But for AI chat to work, Ollama needs to be running! You can see this in the taskbar
I also need help!
If you are good at AI prompts, then DM me (discord: anutays)
Also for those who have a very powerful computer:
You can find and download other models on the Ollama website that are much more powerful,
such as llama3.1:70b (requires about 40 GB of RAM), or other models.
To run a program with a different model, you need to specify it when you run the jar file.
For example: java -jar "Full Path To Jar File" llama3.1:70b
Download: https://drive.google.com/file/d/1yNOqHQ ... sp=sharing
Mirror: https://mega.nz/file/rrYHDD7A#AxSSb1bSZ ... fjJKS0HUUA
I'm currently trying out prompts.