- Temporary local Chat Mode
- Supports a variety of OpenAI models and web-search
- Supports web-search for some models
- Convert responses to HTML or VOICE
- Maintains a log of reponses
- Simple GUI
- Renders HTML from Markdown
- Choose from many themes and colors
- Pop-up Prompt Manager and Options Manager
- RAG for .pdf .md .txt .csv .html documents
requires modules
anthropic
openai
google-genai
ollama
Markdown
ttkbootstrap==1.14.2
groq
Uses Python and tkinter
python3
python3-tk
Ollama (see website to download)
aditional requirements:
text editor (of your choice)
VNC media player
Internet and Internet browser (of your choice)
Tested on Windows and Linux
Use pip or pip3 to install any missing python modules
needed for API access. Consult the vendor for
the correct pip... syntax.
Also you will need API keys from each vendor.
The vendors offer both free and paid tiers.
Here are the vendors, websites, and key labels to use in your environment:
| Vendor | Website | ENVariable |
|---|---|---|
| OpenAI | https://platform.openai.com | GPTKEY |
| Anthropic | https://claude.ai/settings | CLDKEY |
| https://aistudio.google.com | GGLKEY | |
| Ollama | https://ollama.com/settings | OMAKEY |
| groq | https://console.groq.com/keys | GRQKEY |
| Deepseek | https://platform.deepseek.com | DSKKEY |
Windows and Linux have their own ways to set these variables.
Local models are installed using Ollama.
Visit the Ollama website for instructions on installing Ollama and downloading local models.
If you plan to use Voice reading of responses, you will need to install VLC media player.
Use Ctrl-Shift-S to play back the reponses.
Each play-back is saved in a separate file in the application directory.
gemma4:e2b-local
gemma4:e4b-local
groq/compound
gpt-5.4-mini
gpt-5.4-nano
gpt-5-mini
gpt-5-nano
gemini-3.1-flash-lite-preview
gemini-2.5-flash-lite
gemini-2.5-flash
gemini-2.5-pro
rag_gemini-2.5.flash-lite
claude-haiku-4-5
claude-sonnet-4-6
claude-opus-4-6
gpt-4.1-mini
gpt-4.1-nano
gpt-4o-mini
gpt-5
gpt-5.4
qwen3-coder:480b-cloud
qwen3.5:397b-cloud
deepseek-chat
deepseek-reasoner
Note:
for local models obtained from Ollama append the model name with -local
for cloud models obtained from Ollama append the model name with -cloud
the rag- prefix is used to mark the Gemini model used for RAG processing
Only Google Gemini models are supported for RAG at this time.
Enter the model you want for RAG and pre-pend rag- to the model name.
When the 1st prompt is submitted, file will be uploaded and then processed.
When "New" is used (new conversation) the RAG files and objects are removed
from the cloud. Memory for RAG is not preserved across app sessions (except in the log).
The options.ini file contains all of the options settings.
There is a GUI that handles options but can use a text editor also.
- New
Begins a new conversation
To change the system instructions message for the current session,
type instruct You are a .......... assistant
into the Prompt Area first, and then click "New".
- View
Displays the log file you named in Options.
- Text
Opens the current response or selection in your text editor
Set up the name of your text editor in options.
- Html
Converts the current response or selection to HTML and opens it in your default browser.
- Options
launches the Options editing program
- Submit Query
Submits prompt to the current AI Model
Ctrl-G and Ctrl-Enter do that too.
- Web
Toggle the "web-search" tool
NOTE: works with most OpenAI models,
Claude Sonnet, Claude Opus, and groq/compound
- Select temporary Model
Select from models listed in
models.dattext file
Selecting a different model forces a new conversationOn startup the "default" model is always selected
The default model is set in options (options.ini)
- Close
Exit the program. Ctrl-q exits the program without confirmation.
Right-Click in the prompt or response area to get a bunch of useful choices.
At startup, if a previous conversation is detected the user is prompted to either continue or start a new conversation. So closing the app does not terminate a conversation. To start a new conversatioin while the app is running use the 'New' button. Multiple conversations are not preserved anywhere, but will remain in the log until it is purged.
| key | action |
|---|---|
| Ctrl-H | This HotKey help |
| Ctrl-Q | Close Program No Prompt |
| Ctrl-Shift-D | Delete Log File |
| Ctrl-Shift-S | Speak the Currrent Text |
| Ctrl-G | Submit Query (Button) |
| Ctrl-Enter | Submit Query (Button) |
| Ctrl-F | Find text |
| Ctrl-N | Find next text |
| Ctrl-J | Open Selected URL |
| Ctrl-R | Clear Prompt Area |
| Alt-P | Open Prompt Manager (or right-click) |
The prompts directory is for storing prompt manager json files.
In addition, a Prompts Manager can create, update, delete and edit prompts. Access the viewer with Alt-P or with the Right-Click drop-down.
END