Skip to content

README.md

Latest commit

 

History

History
106 lines (72 loc) · 3.31 KB

README.md

File metadata and controls

106 lines (72 loc) · 3.31 KB
Raw

Journals/Blogger Import

Community Contribution

Created by @alanshurafa

Reviewed and merged by the Open Brain maintainer team — thank you for building the future of AI memory!

Import blog posts from Google Blogger Atom XML exports into Open Brain.

What It Does

Parses Google Blogger's Atom XML export format and imports blog posts and comments as thoughts with embeddings. Works with any standard Atom feed export. Blog posts are stored as type: journal thoughts.

Prerequisites

  • Working Open Brain setup (guide)
  • The content-fingerprint-dedup primitive installed (provides the upsert_thought RPC function used for deduplication)
  • Blogger export files — .atom files from Google Blogger
  • Node.js 18+ installed
  • OpenRouter API key for embedding generation

Credential Tracker

JOURNALS/BLOGGER IMPORT -- CREDENTIAL TRACKER
--------------------------------------

FROM YOUR OPEN BRAIN SETUP
  Supabase URL:          ____________
  Service Role Key:      ____________

FROM OPENROUTER
  API Key:               ____________

--------------------------------------

Steps

  1. Export your blog data:

    • Go to your Blogger Dashboard → Settings → Manage Blog → Back up content
    • Download the .atom file
    • If you have multiple blogs, export each one

  2. Place all .atom files in a folder:

    blogger-exports/
├── my-tech-blog.atom
├── personal-journal.atom
└── travel-blog.atom

  3. Copy this recipe folder and install dependencies:

    cd journals-blogger-import
npm install

  4. Create .env with your credentials (see .env.example):

    SUPABASE_URL=https://your-project.supabase.co
SUPABASE_SERVICE_ROLE_KEY=your-service-role-key
OPENROUTER_API_KEY=sk-or-v1-your-key

  5. Preview what will be imported (dry run):

    node import-blogger.mjs /path/to/blogger-exports --dry-run

  6. Run the import:

    node import-blogger.mjs /path/to/blogger-exports

Expected Outcome

After running the import:

  • Each blog post becomes a thought with type: journal and source_type: blogger_import
  • Post titles and publication dates are preserved
  • HTML content is stripped to plain text (line breaks preserved)
  • Blog comments are imported separately
  • Settings and template entries are automatically filtered out

Scale reference: Tested with 3,000+ blog posts across multiple blogs imported successfully.

Troubleshooting

Issue: No entries found in .atom file The parser looks for <entry> tags with kind#post or kind#comment categories. Blogger settings and template entries are filtered out. If your Atom file uses a different schema, check the XML structure.

Issue: HTML tags appearing in imported text The HTML stripper handles common tags and entities. If you see raw HTML in your thoughts, the post may use unusual HTML structures. The content is still searchable.

Issue: Wrong dates on posts The parser uses the <published> tag from the Atom feed. If dates look wrong, check the timezone in your Blogger export settings.