Running into the character limit on complex quizzes when "Create with AI" (Roadmap / MCP question?)
just tested the "Create with AI" feature to build a multi-step, scored lead quiz (a "Which dog fits your lifestyle?" quiz with calculation values and a score aggregation field.
AI generated everything, assigned calc values and even configured the formula inside the Numeric Calculation filed. Awesome!
However, the character limits in the AI prompt modal (2,000 chars / 1,000 chars) are quite restrictive once you want to pass detailed scoring logic, step structures or longer questionnaire content. I had to heavily compress and strip down my context to make it fit.
Are there plans to increase or remove the character limit in the AI prompt builder for users who want to feed richer context? or will the AI Prompt Builder be obsolete when/ if there's an MCP Integration?
If you use directly any AI, you have no rate limits. At least, that has been my experience.
The thing is, it is also more flexible (eg.: you can ask not only to draft the form, but to create the landing, use the created form, thank you message / redirection, send a campaign with the test... everything on the same prompt.
To be honest, I am just expecting the whole thing to bring me a coffee too haha
Jorge de los Reyes did i miss this? are we already capable of connecting e.g. claude?
You didn't miss a native Claude integration. What I meant is that you can already build your own connection today using WPManageNinja's existing APIs and an integration layer you control.
You give Claude (or any other AI tool) the relevant developer documentation and expose only the endpoints and actions you want it to use. I've been working this way for several months and, once it is set up, it feels much like any other integration.
MCP would make the process far more standardized and straightforward, especially around tool discovery and reusable connections, but it is not a prerequisite.
One practical difference is that an MCP server typically exposes a curated set of tools, while a custom integration can work with any documented endpoint you choose to implement and authorize. The APIs already let you build useful workflows today; MCP would make them easier for AI clients to discover and use consistently.
I've recently shared a couple of examples, including this one on using FluentPlayer with AI. I also have another video coming soon with 20+ practical uses across the WPManageNinja ecosystem, including creating FluentCRM contacts and campaigns, among other things.
It is not a tutorial, to be honest, because there are many ways to build this. It is simply an overview of what can already be achieved at the moment.
Jorge de los Reyes watched the video....think i'll have to play a little bit more with AI...😂
Manuel Müller Haha, yes, playing with it is probably the fastest way to understand what is already possible 😄
And just to clarify something that may also help others: the video shows the capability, not one fixed delivery model.
For a one-off project, an agency can run the workflow for the client. But for recurring use, I would normally configure it inside the client’s own environment: their WordPress installation, their API credentials and their preferred AI tool. A small custom MCP server can expose only the approved actions, although MCP is optional because the same workflow can already be built directly with the existing APIs.
The important part is not to give an agent unrestricted access. I would use narrowly scoped permissions, selected endpoints, logs and human approval before anything is published.
So yes: you can build it for clients to use themselves. They would not need to use any speciric chatbot or my API credentials.