Skip to content

Commit 0db6341

Browse files
SimonSchubertSimonSchubert
committed
Add commands
1 parent dc6d62b commit 0db6341

2 files changed

Lines changed: 218 additions & 0 deletions

File tree

‎assets/commands/koboldcpp.md‎

Lines changed: 133 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,133 @@
1+
# TAGLINE
2+
3+
Local AI text generation server and inference engine
4+
5+
# TLDR
6+
7+
**Launch** with a GGUF model file
8+
9+
```koboldcpp --model [path/to/model.gguf]```
10+
11+
**Launch** with GPU acceleration using CUDA
12+
13+
```koboldcpp --model [path/to/model.gguf] --usecuda --gpulayers [35]```
14+
15+
**Launch** with Vulkan GPU support on a custom port
16+
17+
```koboldcpp --model [path/to/model.gguf] --usevulkan --gpulayers [35] --port [8080]```
18+
19+
**Run a single prompt** without starting a server
20+
21+
```koboldcpp --model [path/to/model.gguf] --prompt "[What is the meaning of life?]"```
22+
23+
**Launch in CLI interactive mode** without the web UI
24+
25+
```koboldcpp --model [path/to/model.gguf] --cli```
26+
27+
**Load a saved configuration** file
28+
29+
```koboldcpp --config [path/to/config.kcpps]```
30+
31+
# SYNOPSIS
32+
33+
**koboldcpp** [_options_] [**--model** _model_path_]
34+
35+
# PARAMETERS
36+
37+
**--model** _path_
38+
> Specify the GGUF/GGML model file to load
39+
40+
**--config** _file_
41+
> Load a .kcpps configuration file
42+
43+
**--usecuda**
44+
> Enable NVIDIA CUDA GPU acceleration
45+
46+
**--usevulkan**
47+
> Enable Vulkan GPU acceleration (AMD/NVIDIA)
48+
49+
**--gpulayers** _n_
50+
> Number of model layers to offload to GPU
51+
52+
**--threads** _n_
53+
> Set CPU thread count for inference
54+
55+
**--contextsize** _n_
56+
> Set maximum context length in tokens
57+
58+
**--port** _n_
59+
> Change server port (default: 5001)
60+
61+
**--host** _addr_
62+
> Bind to a specific IP address
63+
64+
**--multiuser** _n_
65+
> Enable multiuser mode with _n_ concurrent slots
66+
67+
**--password** _key_
68+
> Require API authentication with the given key
69+
70+
**--cli**
71+
> Launch interactive command-line interface without starting a server
72+
73+
**--prompt** _text_
74+
> Run a single prompt, print output, and exit
75+
76+
**--benchmark**
77+
> Run performance benchmarking mode
78+
79+
**--flashattention**
80+
> Enable flash attention for improved performance
81+
82+
**--smartcontext**
83+
> Enable smart context handling to reduce reprocessing
84+
85+
**--usemmap**
86+
> Enable memory-mapped file I/O for model loading
87+
88+
**--usemlock**
89+
> Force model to remain in RAM (prevent swapping)
90+
91+
**--ssl**
92+
> Enable SSL for HTTPS connections
93+
94+
**--remotetunnel**
95+
> Enable remote tunnel access for sharing the server
96+
97+
**--sdmodel** _path_
98+
> Load a Stable Diffusion model for image generation
99+
100+
**--noavx2**
101+
> Compatibility mode for older CPUs without AVX2
102+
103+
**--showgui**
104+
> Show the GUI launcher even when command-line flags are used
105+
106+
**--help**
107+
> Display all available commands
108+
109+
# DESCRIPTION
110+
111+
**koboldcpp** is a self-contained AI text generation server that runs large language models locally. Built on top of **llama.cpp**, it provides a bundled web UI (KoboldAI Lite) and supports all GGML and GGUF model formats. It requires no external dependencies and runs as a single executable.
112+
113+
The server exposes an API compatible with KoboldAI and OpenAI formats, making it usable with a wide range of frontends and applications. It supports CPU inference as well as GPU acceleration through **CUDA** (NVIDIA), **Vulkan** (AMD/NVIDIA), and **Metal** (Apple Silicon).
114+
115+
Beyond text generation, koboldcpp supports **image generation** (Stable Diffusion), **speech recognition** (Whisper), and **text-to-speech**, all within the same executable. The bundled web UI offers multiple interaction modes including chat, instruct, adventure, and story writing.
116+
117+
# CONFIGURATION
118+
119+
When launched without arguments, koboldcpp opens a **GUI launcher** for interactive configuration. Settings can be saved to and loaded from **.kcpps** configuration files. Command-line flags override GUI settings when both are used.
120+
121+
Key configuration considerations include **GPU layer offloading** (more layers on GPU means faster inference but requires more VRAM), **context size** (larger contexts use more memory), and **thread count** (typically set to the number of physical CPU cores).
122+
123+
# CAVEATS
124+
125+
Model files can be very large (several GB to over 100 GB) and require significant RAM or VRAM. GPU acceleration requires appropriate drivers and hardware support. Performance varies significantly based on model size, quantization level, and available hardware. The Vulkan backend is more broadly compatible but generally slower than CUDA on NVIDIA hardware. Flash attention requires compatible model architectures.
126+
127+
# HISTORY
128+
129+
KoboldCpp was created by a developer known as **LostRuins** (alias **Concedo**) and first released on **March 16, 2023** as a fork of **llama.cpp** combined with the KoboldAI interface. It was designed to provide a simple, self-contained way to run large language models locally without complex setup. The project grew rapidly alongside the open-source LLM movement, continuously adding features like multi-modal support, GPU backends, and image generation capabilities. It is licensed under **AGPL-3.0**.
130+
131+
# SEE ALSO
132+
133+
[llama](/man/llama)(1), [ollama](/man/ollama)(1)

‎assets/commands/lesskey.md‎

Lines changed: 85 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,85 @@
1+
# TAGLINE
2+
3+
Customize key bindings for less
4+
5+
# TLDR
6+
7+
**Compile** the default key binding source file
8+
9+
```lesskey```
10+
11+
**Compile** a specific key binding file
12+
13+
```lesskey [path/to/lesskey_file]```
14+
15+
**Compile** with a custom output file
16+
17+
```lesskey -o [path/to/output] [path/to/lesskey_file]```
18+
19+
**Print** the version number
20+
21+
```lesskey -V```
22+
23+
# SYNOPSIS
24+
25+
**lesskey** [**-o** _output_] [**--**] [_input_]
26+
27+
# PARAMETERS
28+
29+
**-o** _file_, **--output=**_file_
30+
> Specify the output file for the compiled binary key binding file. Defaults to **~/.less**
31+
32+
**-V**, **--version**
33+
> Display the version number of lesskey
34+
35+
**--**
36+
> Marks the end of options; useful if the input filename begins with a dash
37+
38+
# DESCRIPTION
39+
40+
**lesskey** is used to specify or customize key bindings for the **less** pager. It reads a source file containing key binding definitions and compiles it into a binary file that less reads at startup. The default input file is **~/.lesskey** and the default output file is **~/.less**.
41+
42+
The lesskey source file can contain up to four sections, each introduced by a header line: **#command** for key bindings, **#line-edit** for editing behavior at the search prompt, **#env** for environment variables, and **#stop** to mark end of file.
43+
44+
In the **#command** section, each line maps a key sequence to a less action. Keys can be specified as literal characters, control characters (**^X**), escape sequences (**\e**), or special keys like arrows (**\ku**, **\kd**, **\kr**, **\kl**), page up/down (**\kU**, **\kD**), home (**\kh**), end (**\ke**), and delete (**\kx**).
45+
46+
The **#line-edit** section controls behavior at the search and command prompts with actions like **forw-complete**, **back-complete**, **delete**, **backspace**, **home**, **end**, **up**, **down**, and **abort**.
47+
48+
The **#env** section sets environment variables that less will see at startup, such as **LESS** for default options and **LESSCHARSET** for character set configuration.
49+
50+
# CONFIGURATION
51+
52+
The lesskey source file uses a simple text format. Example **~/.lesskey**:
53+
54+
```
55+
#command
56+
j forw-line
57+
k back-line
58+
\kd forw-line
59+
\ku back-line
60+
gg goto-line
61+
G goto-end
62+
q quit
63+
64+
#line-edit
65+
^A home
66+
^E end
67+
^U abort
68+
69+
#env
70+
LESS = -i -M -R
71+
```
72+
73+
Actions available in the **#command** section include **forw-line**, **back-line**, **forw-screen**, **back-screen**, **forw-scroll**, **back-scroll**, **goto-line**, **goto-end**, **repaint**, **quit**, **search**, and many more. Use the action **invalid** to explicitly undefine a key.
74+
75+
# CAVEATS
76+
77+
In **less version 583** and later, less reads the **~/.lesskey** source file directly without needing to compile it via lesskey. The lesskey binary is considered **deprecated** in modern versions, though it is still distributed. If both **~/.lesskey** (source) and **~/.less** (compiled) exist, newer versions of less prefer the source file. The compiled binary format is not portable across architectures or less versions. The **#env** section can set **LESSOPEN** and **LESSCLOSE** which execute external commands, so lesskey files from untrusted sources should be treated with caution.
78+
79+
# HISTORY
80+
81+
**lesskey** was created by **Mark Nudelman** as a companion utility to the **less** pager, which he developed starting in **1983** as a replacement for **more**. Historically, less required a compiled binary key binding file because parsing a text file at startup was considered too slow. Starting with less version **583** (circa 2021), less gained the ability to read the lesskey source file directly, making the lesskey compilation step optional.
82+
83+
# SEE ALSO
84+
85+
[less](/man/less)(1), [more](/man/more)(1)

0 commit comments

Comments
 (0)