Real Use Case

Use Case #10: Session Archiving to Obsidian

How I built a system that archives every Claude conversation to a searchable Obsidian vault.

W

William Welsh

Author

Jan 11, 2026
6 min read

Use Case #10: Session Archiving to Obsidian

Claude doesn't remember between sessions. But I do.

Every conversation I have with Claude Code gets archived to Obsidian. 237 sessions. 8.6 MB of conversation. Fully searchable.

The Problem

I kept solving the same problems. Different projects, same patterns. But I couldn't remember the exact prompt that worked, or which session had that clever solution.

The Solution

The /archive-sessions command scans all Claude session files, extracts conversations and metadata, converts to Obsidian markdown, tags by project, date, branch, and links related sessions.

How I Use It

Pattern Mining. Search "how did I fix the TypeScript error" across all sessions. Find the one that worked.

Prompt Reuse. That perfect prompt for database migrations? It's in the archive. Copy, paste, adapt.

Learning Loops. Review old sessions. See what worked. See what didn't. Improve.

The Compound Effect

Session 1: Solved a problem. Session 50: Solved it faster because I found Session 1. Session 200: Built a system that solves it automatically.

Each session makes future sessions better. That's institutional memory.

Why Obsidian

Local-first (no cloud dependency). Markdown (future-proof). Graph view (see connections). Full-text search (find anything). Backlinks (discover related sessions).

The Real Value

Last week I searched "supabase RLS policies" in the vault. Found 12 sessions where I'd worked on that topic. One had the exact pattern I needed.

Time saved: 2 hours.

Across a year, with 500+ sessions? This is my second brain for coding.

Try It Yourself

Copy this to set up your own session archiving:

I want to archive my Claude Code sessions to Obsidian for searchable reference.

**Setup:**
1. Where is your Obsidian vault? (path or "I don't have one yet")
2. How should sessions be organized? (by date / by project / by topic)
3. What should be captured? (full conversations / user prompts only / summaries)

I will:
- Create an archive script that reads Claude session data
- Format sessions as markdown for Obsidian
- Add frontmatter with metadata (date, project, tags)
- Set up folder structure for organization
- Create an index file for navigation
- Optionally: set up auto-tagging based on content

Want me to build your archive system?

Archive system built December 2025. Growing daily.

W

William Welsh

Building AI-powered systems and sharing what I learn along the way. Founder at Tech Integration Labs.

View Profile
Share this article:

Related Articles

View all →