Typing MindReviews of typingmind.com

I am learning how to build effective agents. That is a good piece. The plugin system is untennable for me....I have accounts now in semi dark web elements to try to connect to basic plugins...Word...PowerPoint. And after all that...still cannot get the plugins to work. If automated workflow is a goal...then the plugins issue is a problemmatic feature. The graphics plugin...D3??? Whatever that is...is really very low level and I have not been able to create anything useful with it. And...as I read before I bought in...the API costs can get rather steep...depending on which LLM you prefer. I am still happy for the choice to play around with the system...just not able yet to really make it do what I thought I would be doing.

the gemini model is not running well lately. i need to regenerate it multiple times

Was going really well till I just tried to start a new conversation today

instead of telling me that I ran out of space on my cloud subscription It is better to prompt to delete old or to add more space on the go for a fee.

like: a lots of thinks dislike: pricing of cloud sync plans - ridiculous! dislike: Lack of image support

I use this app daily and find it extremely valuable. However, there are a few areas that could use improvement. The file handling functionality is limited - while you can work with files in the app, the ability to download them afterward doesn't function properly. Additionally, the scrolling experience sometimes feels unnatural or inconsistent, which can be frustrating during longer conversations. One more suggestion: it would be better if answers were always displayed starting from the beginning of the response rather than the end, as this would make it easier to follow the AI's complete thoughts. Despite these minor issues, this is still a useful tool.

its been pretty good, but the recent advances with the open ai monthly account interface and with claude, both of which support projects now, I've been shifting away from solely API dev. In particular; the claude system now supports syncing with git, so I can commit and push for versioning and then sync my project in Claude for a very well functioning IDE that has been getting my work done. I would like to see more advancements like this on the TM side, possibly even being able to use your not API with TM. I do like that you're constantly adding new features tho, so I hope to keep using it. I'm excited to try the new Grok API.

I own lifetime access (owned it for approx a year), and I use typingmind alot, mostly via openrouter. (pls, pls pls, for the love of god, make it possible to actually search and filter the list of typingmind models when inputing typingmind api, aswell as set plugin/vision and any other setting individually for each chosen model.) I feel that locking the ability to run a locally self hosted version behind an enterprise price-level subscription is not cool. Let power-users have more control, simple as that. Other points: **Add mcp-server support** - the existing open ai compliant plugin framework is good, but not having native standardized mcp server support sucks. I'm aware this can be done with local js doing some jsdelivr trickery, but an official implementation of this would be very cool, and save alot of headache when making plugins. open ai - mcp software bridge essentially. -- Ability to download files from knowledge base. Once uploaded you can give ai access to files, but you yourself cannot access them, say i want to double check if im giving ai-agent access to what I actually intend? nope, cannot do that. (KB is still in beta). Also I miss more info on how the KB manages the data, does it vectorize it? if so, how? if vectorized, how does it search for and prioritize results? This is important to know in context of how I format and organize the data before upload, aswell as strategy for tagging said data in the KB. Add option to use your own self hosted KB? i.e allow to connect aws S3 API compatible storage. All in all, typingmind is very cool, and nice to use.