DEV Community

Cover image for PlantUMLApp 3.0 - Let's play with AI Multi-Modality
bsorrentino
bsorrentino

Posted on

PlantUMLApp 3.0 - Let's play with AI Multi-Modality

Version 3.0 is out!

New version 3.0 of my PlantUML App for iPad is out with exciting update! 🀩 The new multi-modality feature now lets you transform hand-drawn diagrams into PlantUML scripts with just a pencil ✍🏻 or your fingers πŸ‘†. Take a look πŸ‘€ to this short on YouTube and download it from App Store to support me πŸ‘πŸ».

The Multi Agents Collaboration

In this app I've used Multi-Agents-Collaboration using the LangGraph for Swift framework, applying the process shown in the diagram below:

diagram of solution

As you can see I've used three Agents with different capabilities:

Agent Vision

This agent able to process images, it is skilled on describe a diagram producing a structured output containing also the diagram tipology useful to involve the right Agent for further processing.

This increase the flexibility of system because the image is translated in structured data that can be processed by agents with differents skills and goals

Agent Translator(1)

This agent is skilled on PlantUML sequence diagram. It get the diagram data and translate them in the PlantUML script

Agent Translator(2)

This agent is skilled on PlantUML generic process diagram. It gets the diagram data and translate them in PlantUML script

Conclusion

The Multi-Modality is a capability for enabling AI to interact with and interpret the diverse range of information that humans encounter in everyday life. Unlock such capability can be crucial to amplify the effectiveness of your AI process.
I've started to explore it and, in the next App version, I'd like to introduce a process to auto-correct errors providing to the agents the execution feedbacks.
So stay tuned! In the meanwhile, enjoy coding! πŸ‘‹

References


Originally published at https://bsorrentino.github.io on April 8, 2024.

Top comments (2)

Collapse
 
sloan profile image
Sloan the DEV Moderator

Hey there!

While we appreciate you sharing posts here, please see DEV's Content Policy available in our Terms and be careful not to share content that is too promotional. The terms state:

Users must make a good-faith effort to share content that is on-topic, of high-quality, and is not designed primarily for the purposes of promotion or creating backlinks.

Going forward, be careful to avoid sharing articles that are too promotional in nature. While it's generally okay to promote your product, business, personal brand, etc., just make sure that the content you share has value beyond promotion.

If you balance out any promotional materials with helpful tutorials, walkthroughs, or other information, then your post should fit just fine. But, if you continue sharing heavily promotional material, it's likely that our moderators will downvote your content, making it less discoverable in our feeds and potentially deindexed from search engines. If your content is regularly downvoted by our mods, this will alert our admin, who may take action to remove your content and/or suspend your account.

We hope you continue to enjoy DEV and share content, but please keep the promotional material to a minimum and keep in mind that the community here is made up of software devs who are primarily trying to help out and learn from one another!

Collapse
 
bsorrentino profile image
bsorrentino

Hi @sloan I read the terms, but I'd like want clarify that all I'm "promoting" is available free in open source and my main goal is to share development knowledge.

In this specific case, the app is also available on App Store at a symbolic price (0.99) with license "offer me a beer".

Hope this clarify and is compliant with Dev.to policy