gpt knew my album. claude didn't.
i was talking to gpt about my album and ended up building a memory layer
i'm pretty into music, and i've been working on an album on the side.
i was using gpt and claude to bounce ideas around: lyrics, concepts, references, the whole thing.
but i kept running into the same annoying problem.
i'd tell gpt something about the album, then move to claude and have to explain everything again.
same thing with my projects, ideas, decisions, everything.
eventually i thought, why doesn't my context just come with me?
so i started building innernet.
basically, a memory layer that sits between you and your ai tools, so your context isn't stuck inside one chatbot.
i pitched the idea to a few developer friends expecting them to tell me it was overkill.
instead, they immediately started telling me how they'd use it for their own projects.
that's when i realised this might actually be a real problem.
we're still building it, but the idea is pretty simple:
your ai should remember you, not just the chat you're currently in.
curious if anyone here has run into the same problem.
Why does my AI memory disappear when I switch tools?
I’ve been using ChatGPT, Claude, Cursor and a few other AI tools pretty heavily, and one thing started bothering me more and more.
Every time I switch tools, I basically have to start over.
One tool knows something I told it last week. Another has a completely different understanding of what I’m working on. I end up copying old conversations, pasting context, explaining projects again, or keeping notes somewhere just so the AI can catch up.
And the more I use AI, the more ridiculous this feels.
We already have AI that can reason, code, research and work across huge amounts of information, but the context around me is still stuck inside individual products.
That’s actually the problem we started exploring with innernet.
The idea is pretty simple: what if your context lived independently from the AI tool?
So instead of ChatGPT having one version of you, Claude having another, and Cursor having basically none, your context could follow you across all of them.
We’re still figuring out what the right architecture for this looks like, but I keep coming back to the same thought:
AI should remember you, not the app you happen to be talking to.
Curious if anyone else has run into this while switching between AI tools.
Am I the only one frustrated by AI context not carrying over?
Has anyone else run into this?
I found myself constantly re-explaining context every time I switched between ChatGPT, Claude, Cursor, Gemini, and other AI tools. It feels like every conversation starts from zero.
I'm currently working on solving this problem and wanted to understand your pov.
If you've faced it too, how are you handling it today? MCP, memory files, copy-pasting context, custom tooling, or something else?
Would love to hear your thoughts.