The All-In-One Package
Probably among the best AI implementation GUIs out there; so easy to access, has support for all the platforms, very easy to use and gets updated often.
So many videos, documentation, just a lot of content too to help you get started.
The feature-set is ever expanding and I'm excited for what's going to be coming next. The automation feature is actually crazy, integrating the scrape URL portion to webhooks can lead to crazy integrations like outputting to notes platforms such as Notion databases/Obsidian... like a webscraper on steroids!
I don't particularly understand why there's limits in place for stuff like history/smart prompts/instructions but I highly doubt I'd reach those limits with my use case anytime soon.
The "minds" feature is kinda odd, I don't particularly understand the use case of it. I've tried uploading for example a whole page snippet of code to reference thinking it would take the context of the code into account when prompting, but it kinda just ignored it.
I assume this is meant for more referencing type of use cases where you can upload a whole bunch of information and ask for it back, rather than understand the context. I had much better experience using custom prompts to save snippets then using those; the answers were much better!
Some feature suggestions:
- Some type of basic formatting would be nice. Doesn't have to be full markdown or anything, but a way to at least distinguish between prompt vs text vs code (during input, it does this beautifully on outputs already!). I see a lot of implementations lack this so this might not be feasible but would be so nice to like prompt with some sort of formatting instead of relying on quotations and the like. Especially with dealing with large context sizes.
- Ability to upload files of some sort. PDFs/github repos/etc. It's very cumbersome to do this atm. Would be so amazing if I can just upload an entire repo instead of copy pasting each file into separate custom prompts.
10/10, will be upgrading to Tier 4 momentarily :]
Felipe_TriploAI
May 9, 2024Hello James! I really appreciate your support and feedback.
I agree with pretty much everything you mentioned. We recognize there's always room for improvement, and we're actively working on it :)
I genuinely enjoy helping our users and providing content that empowers you to make the most of Triplo AI. You can look forward to more videos, detailed documentation, use cases, and more. I wish there were more hours in a day to do everything, but I must say Triplo has been a huge help too, hahaha.
Here are a few quick points I think are important to address:
- I'm also a big fan of automations. It's great to hear that you love it too! It's incredibly powerful!
- Some of the limits are in place to prevent abuse. For example, history isn't implemented yet (and probably won't be), and others intentionally have high thresholds to make our users feel comfortable.
- Training, this is important. It's intended to function as a knowledge base for context, facts, or references you trust. I don't see it working well with long code, for example. It's more suitable for references such as documentation, papers, news, or blog posts, etc. It was designed to cover topics where mistakes are unacceptable, and you need to ensure the correctness of the content.
- Let me consider formatting. I'm not sure how to implement it, but we'll discuss it.
- Files/URLs will definitely be integrated into Triplo AI eventually, starting with Training, and then we'll see.
Once again, thank you for your engagement and support. It means the world to us.
Take care,
Felipe and the Elbruz Tech Dream Team