Notes 
KoboldCpp is a self-contained local AI inference tool with a built-in web UI. It runs GGUF text, image, and speech models with no installation beyond a single binary, primarily aimed at creative writing and roleplay workflows.
Reviewed by Marcus Holmberg
KoboldCpp
github.com/LostRuins/koboldcpp No ratings yet. Be the first.
Rate KoboldCpp: No account needed
Threat level
Hardened
A hardened pick. Worth the effort once you have chosen to shrink your footprint on purpose. Enough for most people. Threat levels
Our take
Where Ollama optimizes for developer API use, KoboldCpp leans into the browser UI and creative writing features: character cards, story memory, chat templates, and image generation through a built-in Stable Diffusion backend. Everything runs locally and privately. The interface is dense and the configuration surface is wide, which rewards power users but can overwhelm newcomers. Skip it if you want a clean API or a simple chat box; reach for it if you want deep control over text generation and a capable all-in-one local setup.
GitHub at a glance
LostRuins/koboldcpp AGPL-3.0 gemmaggmlggufkoboldaikoboldcpplanguage-modelllamallamacppllmmistral
Stars
12,145
+62 this week
Forks
807
108 watching
Last commit
today
healthy
Latest release
v1.124.2
4d ago, 100+ total
Star history
Issues
67% closed
528
open
1,095
closed
+23 / -17
opened / closed, last 4 weeks
Opened vs closed, last 12 weeks
Commits, last 26 weeks
0 commits, active 0 of 26 weeks
443 contributors since Mar 2023
Languages
- C++ 92%
- C 2.9%
- Python 1.7%
- Other 3.4%
Updated daily from the GitHub API. Star tracking since Jun 22, 2026.
KoboldCpp alternatives
Lumo Proton's hosted AI assistant with zero-access encryption of saved chats and a strict no-logs policy.
Duck.ai DuckDuckGo's private AI chat that lets you talk to multiple models with no account and no training on your prompts.
Euria Infomaniak's Swiss-hosted AI assistant that pledges no training on user data and operates under GDPR and Swiss data law.
Apple Intelligence Apple's AI. On-device where it can, with verifiable Private Cloud Compute for the rest.
Skales Skales is a local AI desktop agent that runs on your computer with any AI provider you choose. All conversations and data stay on your machine, and you control which model and which API keys you use.
Ollama Ollama lets you download and run large language models locally via a simple CLI and REST API. Supports a growing library of open models including Llama, Mistral, and Gemma on Windows, Mac, and Linux with no data sent to the cloud.
KoboldCpp license, in plain English
AGPL-3.0 Strong copyleft · network
Free to use, even commercially. Changes must be published under the same license, and running a modified version as a network service counts as distributing it.
Permits
- Commercial use
- Modification
- Distribution
- Patent use
- Private use
Requires
- Disclose source
- Network use is distribution
- Same license
- State changes
- License and copyright notice
Does not provide
- Liability cover
- Warranty
Why it matters: The network clause is the point. Anyone who runs a modified version as a hosted service has to publish those changes, so the code handling your data stays inspectable. This is why privacy-first projects reach for AGPL.
Plain-language summary of the project's license, not legal advice. Read the full text for the exact terms.
* Average ratings will show on this page once the threshold of 5 ratings is reached.