Support

Memory Import

Bring the memory you've already built in ChatGPT, Claude, Cursor, or Mem0 into Tempreon in a few minutes.

Overview

You've probably already taught your AI tools a lot about yourself. Memory Import brings that existing memory into Tempreon so you don't start from a blank slate — and once it's in Tempreon, it travels to every tool you connect, not just the one that learned it.

It's free on every plan, takes about 2–3 minutes, and ends with a reveal of your updated Kernel.

Supported Sources

ChatGPT, Claude, Cursor, and Mem0 — plus an "other" path for any AI tool that can read its own memory.

Before You Start

The import runs through your connected AI tool, so the tool you're importing from needs Tempreon connected as a Bridge first. If you haven't done that yet, start with the setup guide for your tool in Bridges.

How to Import

  1. Go to your Bridges page and find the import section.
  2. Pick the tool you're importing from. Tempreon generates a tailored bridge prompt.
  3. Copy the prompt and paste it into a chat in that tool. The tool reads what it already knows about you and stores it into Tempreon — it'll ask whether you want to review items together or just proceed.
  4. Watch the live counter in the Tempreon modal tick up as items land.
  5. When your tool says it's done, click "I'm Done" to finalize.
  6. The completion screen shows what your import added — and your updated Kernel.

Good to Know

  • Duplicates merge. Importing something Tempreon already knows won't create copies.
  • Undo works. A whole import run can be undone if it wasn't what you wanted.
  • At capacity, nothing is deleted. If an import pushes you past your plan's memory capacity, the least-recently-used memories are set aside into a recovery bin. Restore them all by upgrading, or export your data anytime — exports include set-aside memories.

Troubleshooting

The counter stays at 0. Your AI tool usually hasn't approved Tempreon's tools yet. Check for a permission prompt in the tool, confirm Tempreon is enabled in that conversation, then send the prompt again.

My tool says it can't see its own memory. Some tools gate memory access behind their own plan or settings (for example, ChatGPT memory has its own toggle). That's a setting in the source tool, not in Tempreon.

It finished but my Kernel looks the same. Kernel lines are earned, not copied — your imported memories are searchable immediately, and new Kernel conditionals appear as the system distills them.

Try Saying...

  • "What did my import add?" — Your AI can summarize the new records
  • "Search my imported memories for [topic]" — Everything is searchable right away