Desktop AI chat for Ollama models
Mostly for Linux (might work on Windows)
GUI desktop AI client for conversing with Ollama's downloadable Models
Of course you will need the Ollama runtime
download from Ollams.com
Suggest 32G memory and a core i9 CPU, or GPU.
| Button | Action |
|---|---|
| Clear | Starts new conversation |
| Export | Transfers response markdown to HTML in system default browser |
| View log | View log text in the response pane |
| Submit | Send query to the AI (Ctrl-G) |
| Close | Exit application |
| Keys | Action |
|---|---|
| Ctrl-H | This help message |
| Ctrl-F | Find text |
| Ctrl-N | Find next |
| Ctrl-Q | Quit App |
| Ctrl-G | Execute AI request |
| Ctrl-O | Open options in text editor |
| Ctrl-D | Delete Log |
| Alt-Ctrl-C | Copy Code in Markup |
# options.ini
# -----------
# llama3.1:8b
# gemma2:9b
# qwen2.5-coder:14b
# deepseek-coder-v2:16b
#
model=llama3.1:8b
role=you are a helpful coding and technology assistant
log=on
font1=Cascadia Code
fontsz1=11
font2=DejaVu Sans Mono
fontsz2=10
height=150
color=#8888ff
editor=mousepad
Options are pretty self explanatory.
"color" refers to text color of the response area.
"height" refers to the height of the prompt area.
If you've set up the "editor" option, then you can
edit the options.ini file by hitting Ctrl-O.
On exiting the program, the conversation is not ended (deleted). Upon starting the program you are prompted to either start a new conversation or continue the previous one.
The conversation is temporary and stored in a file called 'conversation.json'
The program only saves each current conversation. Past conversations are not labeled and stored, but would be reflected in the 'log' until you delete the log (Ctrl-D).