What
Extend tools/legacy_importer.py to support importing conversation history from:
- Google Gemini (via Google Takeout)
- Perplexity AI
- Character.ai
Currently supports Claude and ChatGPT exports.
Why
The "Session 000" seeding story is one of Adam's strongest pitches — import your old conversations and your AI starts with real history instead of blank. More import formats = more people who can start with a populated Vault on day one.
What exists
tools/legacy_importer.py already handles:
- Claude JSON exports → Vault session format
- ChatGPT JSON exports → Vault session format
The pattern is: parse export → extract sessions → write as Vault-compatible markdown → optionally ingest into neural graph.
What to add
Each new format needs:
- A parser for that platform's export format
- Mapping to the Vault session schema (date, turns, topics)
- Test fixtures with sample export data (synthetic, not real conversations)
- Documentation in the script's help text
Gemini is highest priority since Google Takeout exports are standardized.
What
Extend
tools/legacy_importer.pyto support importing conversation history from:Currently supports Claude and ChatGPT exports.
Why
The "Session 000" seeding story is one of Adam's strongest pitches — import your old conversations and your AI starts with real history instead of blank. More import formats = more people who can start with a populated Vault on day one.
What exists
tools/legacy_importer.pyalready handles:The pattern is: parse export → extract sessions → write as Vault-compatible markdown → optionally ingest into neural graph.
What to add
Each new format needs:
Gemini is highest priority since Google Takeout exports are standardized.