HuggingFace Integration Guide (Lightweight)
Philosophy: Keep Boring Core light using HuggingFace's standard tools (
git,git-lfs) for free cloud storage.
This guide explains how to backup and share your .boring-brain files on HuggingFace Datasets.
Why HuggingFace?
- Free: Unlimited storage for public datasets.
- Git-Native: Fits perfectly with Boring's GitOps philosophy.
- LFS Support: Excellent support for large files (Vector DBs).
Prerequisites
- Create a HuggingFace Account.
- Install Git LFS:
git lfs install. - Set up SSH Key or Access Token.
Workflow
1. Export Brain
First, export your Agent's knowledge into a single file:
2. Create HuggingFace Dataset
- Go to HuggingFace New Dataset.
- Set a name (e.g.,
my-username/boring-knowledge). - Create Repository.
3. Upload (Git LFS)
# 1. Clone the repo you just created
git clone https://huggingface.co/datasets/my-username/boring-knowledge
cd boring-knowledge
# 2. Enable LFS tracking for .boring-brain files
git lfs install
git lfs track "*.boring-brain"
git add .gitattributes
# 3. Copy and Commit
cp ../my-knowledge.boring-brain .
git add my-knowledge.boring-brain
git commit -m "feat: update brain knowledge dump"
# 4. Push
git push
4. Download & Import (Teammate's View)
Your teammate can now easily fetch the knowledge:
# Download
wget https://huggingface.co/datasets/my-username/boring-knowledge/resolve/main/my-knowledge.boring-brain
# Import (Merge Mode)
boring brain import my-knowledge.boring-brain
Automation Tip
You can wrap these steps into a simple Shell Script (upload_brain.sh) in your project.