Folks, are there GenAI tools that can be used like ChapGPT prompt? Basically, chapgpt seems to now require login and i don’t want the thing to profile me. I’d rather train a foss tool.

  • FrankLaskey ( FrankLaskey@lemmy.ml ) 
    link
    fedilink
    English
    arrow-up
    2
    ·
    9 個月前

    I would check out Open WebUI which can be self-hosted via docker etc and configured with any OpenAI compatible endpoint so you can use a service like OpenRouter to run nearly any LLM remotely. Most of the open weights ones like Qwen 3 or Kimi K2 Thinking are great and cost pennies per inquiry and can be configured with Zero Data Retention (ZDR) so your data is not recorded. You could also use something like Ollama to run local LLMs if you want even more privacy and have the hardware (typically a modern Nvidia GPU with at least 16-24 GB of VRAM).

  • Auster ( Auster@thebrainbin.org ) 
    link
    fedilink
    arrow-up
    1
    ·
    9 個月前

    Dunno how to train a database model or however it’s called, but using Termux + Ollama + qwen2.5:3b, it helps as much as I’d expect for a (to my knowledge) non commercial tool. Also since you can pull other models/whatever, maybe you can provide your own too.

  • Sims ( Sims@lemmy.ml ) 
    link
    fedilink
    arrow-up
    1
    ·
    9 個月前

    Framework:

    Advanced generic agent: Agent-zero (needs setting up local/external endpoints + invest time in optimizing)

    Easy setup/integration with std linux: Newelle (flatpak, and downloads own dependencies/extensions. You choose)

    Working advanced chat/agent abilities out of the box: cherry-studio (flatpak, default free glm4.5/qwen models can be changed)

    Local endpoints: llamacpp, ollama, vllm, transformer (huggingface) or depending on your hardware. Another option is gpt4free that are a tool/library that collects free chat on the net and set them up as a local endpoint. Aihorde is also great.

  • For Android, there are plenty of local AI clients such as ChatterUI, PocketPal etc. Just download a suitable gguf model from huggingface and use it. If you have 8GB+ RAM, you can easily run 3B models.

    Edit: Try to find iMatrix quantized gguf models. They preserve better quality in in smaller size and runs bit faster.

    • trilobite ( trilobite@lemmy.ml ) OP
      link
      fedilink
      arrow-up
      1
      ·
      9 個月前

      None of these are on fdroid. Only Ollama chatta is there but was updated 16 months ago. Hugginface looks interesting. Are the model you download from there safe?

      • HoleSailor ( HoleSailor@feddit.org ) 
        link
        fedilink
        arrow-up
        1
        ·
        edit-2
        9 個月前

        I use ChatterUI (available on Github) because I found it faster and smoother than the rest of the clients available for Android.

        Yes, the gguf models on Huggingface are totally safe. They will run locally in your phone. You don’t need Internet connection to run that model. So, it is safe. No data will ever leave your device.

        Edit: There is only one client available on f-droid: Maid But it never worked in my device, unfortunately.