Previously, I made a CKBGPT using OpenAI’s MyGPT feature, but since it required a paid subscription, an alternative approach was explored to make CKB-related knowledge more accessible. With DeepSeek’s API now being much more affordable, repository contents from the following GitHub organizations were extracted and processed:
- ckb-devrel
- sporeprotocol
- utxostack
- fi5box
- Nervos Network’s GitHub (pinned repositories and the Fiber repository)
To make the extracted content more LLM prompt-friendly, the open-source tool gitingest was used. Then, Cherry Studio was leveraged to embed the contents as a knowledge base.
The resources are shared in case they are useful to others:
- Extracted Codebase: For further development Download
- Embedded Knowledge Base: For direct use with Cherry Studio Download
How to Use with Cherry Studio
For those who want to explore the knowledge base in Cherry Studio, here’s how to set it up:
- Download Cherry Studio
- Import the Database:
- Go to Settings → Data Setting → Data Backup and Recovery
- Import the downloaded knowledge base
- Model Configuration:
- Refer to Cherry AI Docs for model-related settings
Then, enjoy!
Future Plans
There are thoughts about integrating this setup with RAGFlow or others to develop a CKB Bot, making it even easier to access information without requiring any local setup.
Any feedback or suggestions are always welcome!