About ContinueVault

One founder in Illinois, USA, who got tired of re-explaining his own project to an AI every time a chat hit the wall.

ContinueVault saves the work you do across Claude, ChatGPT, Gemini, Grok, DeepSeek, Perplexity, and Le Chat into one encrypted vault, makes it searchable, and lets a fresh session in any of them pick up the thread. It is a product from ContinueVault LLC, a small company started in 2026 and based in Illinois, USA.

Why this exists

I was deep in a long-running project in ChatGPT. Tweak after tweak, hour after hour. The screen got slower and slower. So I opened a new chat: fast again, and it remembered nothing about the hours we had put in. I pasted the old chat into the new one. Slow again. I tried Projects. I tried uploading a document of the whole conversation. All the things everyone tries.

Eventually enough work was buried in dead chats that I had to understand what was going on. What I found is that a new session is a spotless mind. The model knows what it was trained on; it knows nothing about you or anything you have ever discussed with it. The built-in “memory” features are a few pages of notes and a keyword search over recent chats. Useful, but not the thing I wanted.

The thing I wanted was simple. Open a new chat, type continue, and have it pick up where we left off: what we tried, what we decided, what was still open. Stop answering the same questions over and over.

So I built it. A lot of experimenting, with the AI itself prodding me about what it needed in order to get back to full speed. The answer turned out to be making every past conversation searchable in a way the model could reach into on its own, and trusting it to go get what it already knew.

Then it got weird, in a good way. The same infrastructure worked from other AIs too. All my ChatGPT history was suddenly available in Claude, and vice versa, and then in Perplexity and Claude Code and the rest. That was never the goal. It was a “I wonder if I can” that worked.

Once it worked, it was obvious other people hit the same wall. So it became a product. The longer version of this story is here.

Who is behind it

Jeremy Waldrop, founder.

Systems engineer with more than 35 years building software that has to work: electronic railcar braking and control systems (five US patents), then two decades of embedded and IoT firmware for field-deployed products, trusted by companies like Ford and Verizon.

ContinueVault is what he built after hours, first for himself, because the tools he was paying for kept forgetting his work. The embedded background is why it is encrypted by default, keeps a minimal data surface, and takes no shortcuts on reliability.

The rest of the team is a group of AI agents running on Claude, ChatGPT, and Codex, coordinated through an issue tracker, and working with ContinueVault as their memory. When you read that the product is used to build itself, that is what it means: yes, where the team thinks.

How we work

Your conversations are intensely personal, so the product is built around that. Each vault is encrypted per user; the keys live only in a process that is serving your authenticated requests and die when you go idle. No scheduled jobs read your history, no support access without your explicit, time-limited consent, and nothing is ever used for model training.

A few things we have deliberately said no to: tracking cookies, fake pricing anchors, and launching before a version is ready. Free is the whole product on your last 7 days; Pro is $19/mo with your full history, and founding members keep that price.

Contact