Brief overview
🔍 Added a small feature to zoom in using mouse selection. Handy for looking at deep branches #ChatCircuit
👉 August 22, 2024
Re-run all nodes in a branch
Chat Circuit now makes it possible to re-run a branch of your conversation with LLM with a different prompt. It supports all local LLMs running on @ollama
💾 👉 August 6, 2024
Generate Follow up questions
Implemented this idea in chat circuit. Here is a quick demo of the application along with generating follow up questions using #LLM August 20, 2024
Zoom in/out
🔍 Added a small feature to zoom in using mouse selection. Handy for looking at deep branches #ChatCircuit
👉 August 22, 2024
Minimap Support
#ChatCircuit Added a mini-map with the help of Sonnet 3.5 in @poe_platform.
Would have taken me days if not weeks to do it without any help. 🙏
~ 99% of code is written by Claude September 25, 2024
Export to JSON Canvas Document
Added option to export to #JSON Canvas document that can be imported by any supported application like @obsdmd / @KinopioClub
👉 September 26, 2024
Multi-Branch Conversations Create and manage multiple conversation branches seamlessly.
Contextual Forking Fork conversation branches with accurate context retention.
Save and Load Diagrams
Undo and Redo
Zoom and Pan
Re-run nodes in a branch
It is possible to re-run all the nodes in a branch after changing the prompt it any node in the list.
To run this application, follow these steps:
Generate models configuration file
ollama list | tail -n +2 | awk '{print $1}' > models.conf
Install dependencies
python3 -m pip install -r requirements.txt
Run application
python3 main.py
The LLM models available are loaded from models.conf
in the current directory
See models.conf.example
The default model is the first one in that list
You can also run this command to generate the models.conf
file
ollama list | tail -n +2 | awk '{print "ollama_chat/"$1}' > models.conf
Note: If models.conf is not found, the application will use a default set of models.