I have a new workflow I’m excited about:
- Have a productive chat with an LLM
- Ask it to output a clean Markdown summary of the key points
- Include YAML frontmatter for Obsidian
- Feed that into an ingestion flow that files it into your vault
I often used to lose valuable insights like this.
The same idea works for project work too:
- Do the work in a dedicated Codex or Claude Code project chat
- Ask that chat to generate an Obsidian-ready note
- Ingest it into your vault as durable documentation
Why this matters:
Chat context is ephemeral.
Notes are durable.
Instead of losing useful thinking in long threads, you turn it into structured, searchable knowledge:
- Decisions
- Insights
- Progress logs
- Next steps
- Reusable context
Bonus: you can use Codex or Claude Code to write into your Obsidian vault.
This is especially handy if you run out of credits in one or the other.
Over time this builds real long-term leverage. Your notes stop being random fragments and start becoming a compounding memory system.
This is super easy to set up.