Welcome To KoboldCpp

These repositories contain resources related to help newbies get started with KoboldCPP.

First - you need to obtain KoboldCpp. You can do so by downloading it from the KoboldCpp Github releases Linux, Windows and MacOS are supported - pick the correct binary for your OS.

Next, you'll need to get a GGUF model. In the kcppt repo there are predefined templates for some popular models, loading them into koboldcpp will automatically download and load all the correct files required. You can also find other GGUF models by searching around Huggingface. Finally, launch KoboldCpp, and it should open a Web UI in your browser. Enjoy!

Other repos here:

newbie templates - Contains premade KoboldCpp quick launch templates curated for newbies.
popular templates - Contains premade KoboldCpp quick launch templates curated for popularity.
kcppt - Contains miscellanous premade KoboldCpp quick launch templates, links should be permanent. Some may be outdated.
tts - Contains text to speech models for KoboldCpp
whisper - Contains voice recognition (transcriber) models for KoboldCpp
imgmodel - Contains some sample stable diffusion models that work in KoboldCpp (many more work too)
mmproj - Contains multimodal projectors (mmproj) for Vision models to use in KoboldCpp (image recognition)
music - Music generation models
gguf dumps - Miscellaneous GGUF models

If you ever get lost or have questions, check out the Wiki at https://github.com/LostRuins/koboldcpp/wiki or come to our discord at https://koboldai.org/discord