Overhaul: Cloud Sync
C
Chris
Syncing is not inuitive and its not affordable
If you sync items (chats, prompts, plugins, etc) to the cloud, then later uncheck items to be synced, like chats, there is no _easy_ way to keep them in TypingMind. You need to delete them from the cloud that will then delete them from TypingMind. So you would have to manually export the chats, then reimport them if you want them in TypingMind.
That's not how syncing is suppose to work.
TypingMind Cloud pricing is insanely high.
C
Chris
📝 Bump
Ted H
I just went through this scenario yesterday, where I went through the TypingMind PWA on my desktop, and deleted most of the stored chats there after backing them up to my local machine, in order to get below the 10MB limit. I needed to get below the 10MB limit so that I could sign back in on my other devices and get the rest of the interface (API keys, plugins, etc) to sync. At this point, any chat that wasn't on my local machine but was in the cloud was deleted. But I was able to bring the cloud storage below 10MB, and able to sync the other devices, so in general it was successful.
It would be helpful if data generated by the user (chats, files, images, etc) could be synched separately from the necessary infrastructure items (API keys, plugins, settings, licenses). As of today, nothing will synch if the limit is reached. Separating these synch items and allowing those smaller essential items to synch even when the chats and attachments are over the limit would alleviate some of the bottleneck.
Allowing a portal to use local storage or our own cloud storage for synching our own data that doesn't include TypingMind settings as well as for RAG use would be a big relief and drive my enthusiasm for TypingMind way up again.
Stephan Fischer
These really need addressing ASAP It renders the application useless for professional use, unless you want to pay the additional exorbitant cloud fees.
C
Chris Cameron
Relevant for this - The price to increase the knowledge base character limit is high. $10/month feels like a lot of money for a 1 million character limit. The reason I want to use RAG is because I have many documents, but the limit
for RAG
is still under the context size of some models.