Agentic coding workflows
-
@Christoph-Hart said in Agentic coding workflows:
But how and why did we get to the point of writing code using AI
Simple: the latest iterations of the SOTA models make AI coding on par with a reasonably well skilled developer.
It is the begining of the end then :)
For those who think AI can't create good plugins, train your AI yourselves, and eventually this prophecy will finally come true :)
-
@JulesV said in Agentic coding workflows:
isn't this technically vibe coding?
Vibe coding is where you prompt the AI and accept the output. When it doesn't work you give it the error message and accept the new output. You don't look at the code, you don't understand the code, and you don't care about its quality.
What we're doing it prompting the AI with a problem. We're examining and understanding its output. We ask for considered changes, or we make the changes manually (although I find I'm doing less of that). Then we test thoroughly and go back to the AI if it needs further changes. The code quality is maintained.
@JulesV said in Agentic coding workflows:
Saying that AI can't produce good plugins while simultaneously training it to do so is, in my opinion, wrong
AI can already produce good plugins, that ship has sailed. But they are not necessarily interesting plugins or widely useful. They are more specific and tailor made for individuals. Although there is no reason why you couldn't use AI to generate a saleable product - I'm sure lots of people already are.
But we are not using AI to create an entire HISE project (yet). We're using it as a tool to help our development process. It's like if you hire a developer and you tell them what you want, you don't write the code yourself, you just steer the developer in the direction you want to take the project and they do the actual grunt work.
@JulesV said in Agentic coding workflows:
Wouldn't it be much better to focus on analytics and DSP topics, as I mentioned above?
DSP and Look and Feel is probably what I will mostly use it for within my HISE projects. But I'm using it a lot for adding features I want in HISE itself, or fixing bugs. I'm sure Christoph is going to be overwhelmed with all my pull requests :) We'll need to train a Christoph AI to manage them.
Free HISE Bootcamp Full Course for beginners.
YouTube Channel - Public HISE tutorials
My Patreon - HISE tutorials -
Nobody is expecting that AI will produce interesting plugins on its own - the same as it will never produce interesting music with any artistic value ever.
It‘s the middle layer that handles the transformation from a human having an idea and vision to a plugin that is completely being revolutionized at the moment - this is where HISE sits and ignoring this paradigm shift would be suicidal for a framework like HISE which is why I‘m pulling 14 hour days since about 3-4 weeks to figure out the best way going forward. The way Opus 4.6 interacts with the current toolset I‘m developing is uncanny though and goes far beyomd simple LAF stuff - it eg. oneshotted the transformation of a fairly complex user preset model to the custom automation handler then Ralphed it‘s way to find out that setting processorIds will make the control callback not fire - with proper guidance and planning, which ties back to daves distinction between vibecoding and AI assisted development.
-
@Christoph-Hart said in Agentic coding workflows:
It‘s the middle layer that handles the transformation from a human having an idea and vision to a plugin
MCP integration so awesome!
Please allow it to add modules to the module tree
-
-
@Bart I haven't tried it but looks like that may be in the works: https://github.com/christoph-hart/hise_cpp_raw_template/blob/main/AGENTS.md
-
@HISEnberg said in Agentic coding workflows:
@Bart I haven't tried it but looks like that may be in the works: https://github.com/christoph-hart/hise_cpp_raw_template/blob/main/AGENTS.md
Thanks HISEnberg.
It does appear like its on the menu
Testing last week I found it can add Hise modules within one of the projects files, but couldn't directly add the modules to the current Hise project it's interacting with.
Also found it gets in to a loop of shutting down and re-opening Hise when the AI needs to create a new project. It seems not able to create new Hise projects, nor load xml's yet.
These few limitations almost make it useable for me atm, but will test again soon
-
@Bart FYI I'm working on a TUI app at the moment that allows you to remote control HISE and perform all kinds of actions - the same command set is also exposed as an CLI interface for the LLM to use (and you can use the TUI app as a "log viewer" to trace back what the LLM did).
This will include a feature complete builder mode with full undo support where you can add / remove modules and change parameters from the command line:
-
@Christoph-Hart excited to see what is in the works here!
-
@HISEnberg haha yes that is some funky stuff, my favorite part is how I replicated the module tree in ASCII art:
I've also pulled in the entire setup.hise.dev stuff into this app - this bypasses some of the quirks of the website approach - now people just have to download this self-contained binary TUI app and type in
/setupand it will perform all the steps from the setup workflow.It also acts as logger for monitoring what the LLM does - if you have the TUI open and the LLM calls
hise-cli -builder add Samplerthen this shows up in the chat history with an "LLM" tag.
I'm probably moving all the runtime functionality out of the MCP server as soon as the hise_cli is ready so the MCP becomes a read-only documentation / explorer tool that sits on the HISE doc server - currently it needs to be run locally so it can connect to the HISE Rest server, but since the Hise-cli is doing the same thing too it's easier to separate these tasks.
The MCP server also integrates with the new doc website and uses a small embedding model and a vector database for looking up docs & code examples - I evaluated different approaches, but that gives me the advantage of checking the quality of the MCP search.
