kgnkcbs/ai-obsidian-n8n
0
AI Obsidian Automation for Hugging Face Spaces
This Space runs n8n on a Hugging Face server and exposes two public endpoints:
POST /webhook/ai-notlarGET /webhook/health
The workflow accepts voice text and clipboard text from iPhone Shortcuts, sends the note to Claude, Gemini, and ChatGPT, then writes the result into your GitHub-hosted Obsidian vault.
What Changed
This package is now server-ready:
- no hard-coded Hugging Face username in the
Dockerfile - no hard-coded GitHub username in the workflow
- no mandatory manual n8n credential setup for OpenAI, Anthropic, Gemini, or GitHub
- runtime config comes from Hugging Face Space Variables and Secrets
Required Hugging Face Variables / Secrets
Copy values from hf-space.env.example:
GITHUB_OWNERGITHUB_REPOGITHUB_TOKEN
SPACE_HOST is already provided automatically by Hugging Face Spaces, so you do not need to add it manually.
Optional AI provider secrets:
OPENAI_API_KEYANTHROPIC_API_KEYGEMINI_API_KEY
Optional model variables:
OPENAI_MODELCLAUDE_MODELGEMINI_MODEL
Deploy
- Create a new Hugging Face Space with
SDK = Docker. - Upload all files from this folder.
- Add the variables/secrets listed above in Space Settings.
- Build the Space.
- Open the Space URL and create the first
n8nowner account. - Import
n8n_workflow_hfspace.jsoninton8n.
Prepare The Workflow
Run this once before uploading if you want the workflow converted to the server-ready version automatically:
powershell -ExecutionPolicy Bypass -File .\prepare-hfspace.ps1That command updates n8n_workflow_hfspace.json in place and keeps the original as n8n_workflow_hfspace.original.json.
