Guide · Migration
Take your conversations between AI assistants
Switching assistants shouldn't mean starting from zero. With a neutral export, your best threads move with you.
Why "import this chat" doesn't exist
Assistants are walled gardens: each keeps its own schema and offers no import button for another platform's history. That is why migration is a two-step process — export on the old side, standardize in the middle — instead of a single button.
Step 1 — Export from the source
Pull your history from any source: a ChatGPT ZIP, Claude JSON export, Gemini Takeout, DeepSeek archive, or even multiple files at once. Names and shapes differ, but the analyzer's universal parser handles them all.
Step 2 — Standardize in the middle
Drop everything into the analyzer to get one normalized view of every platform at once: search across all of them with one query, spot your best threads, and extract the custom instructions you want to take along. It turns a pile of incompatible formats into one browseable archive.
Step 3 — Export in a portable format
For human use, export clean Markdown transcripts — the universal format you can open anywhere, paste into any new assistant as context, or drop into a wiki. For machine use, take the normalized JSON mirror. Either way the content is now assistant-agnostic.
Step 4 — Carry the instructions over
The part people forget: your old curator personality or code style lives in custom instructions and system prompts. Read them from the analyzer's Instructions view, then paste them into the new assistant's settings. That is where the actual "personality" of your workflow is stored.
Questions
Can I import my ChatGPT history into Claude or Gemini directly?
No platform offers a one-click import for transcripts today. The practical path is standardizing: export everything, then keep a normalized, portable archive you can dip into anywhere.
Do I need to pay to migrate?
No. Export, parsing and export-to-Markdown/JSON all run free and locally in the analyzer.
Will my custom instructions transfer?
Not automatically. The analyzer extracts them from system messages so you can copy them into your new assistant's settings by hand.
What is the lowest-effort migration?
Export → analyze → export clean Markdown per conversation. If you only need the content, plain transcripts are the universal format.