Understanding AI Usage Costs and Credit Consumption per Task in Publish Platform

·Aug 28, 2026 04:36 PM·

Hey guys, I tried to find the documentation on AI usage, but I’m trying to see how or what the cost are per request or task in publish. Does everything from brief to design and everything else cost credits, trying to gauge this so I’m being responsible with my projects. Thanks in advance for the info.

6 comments

· Sorted by Oldest

Matt J.
·
·
Issac A. not entirely sure we'd share costs - I'll ask or someone can step in to answer that directly but I do know that at the end of the day, we're monitoring AI usage and are willing to adjust accordingly. I think the last thing we want to do is block work from happening because of token usage - we just have to be reasonable about eating costs. Its why we deliberately avoided having a token counter or usage meter. AI usage should be a non-issue, something you dont have to think about or worry about. The AI team have massive dashboards that are tracking various metrics and cost is one of them - a small percentage of new users are hitting the limit on the free plan right now, and even less people are hitting the limits on the Pro 5x or Pro 20x plans. In other words, push it and let's adjust. We're listening!
Matt J.
·
·
also should be said that Relume is indeed eating some costs on the free plan with AI - we're willing to do this so that users have a solid opportunity to test drive.
Issac A.
·
·
Uhhh, seems like I might have communicated this poorly again. All I want to do is know what features generally use AI credits, like prompting the brief, or generating a page. I don't want to know the API costs to you, l lol
Matt J.
·
(edited)·
ahhh - hope its not me, not comprehending your question correctly haha. I'll be honest, I try to read the questions a couple of times to make sure but yeah. So yeah, a lot of things touch AI - we're in the tens of billions of tokens used up already across the system. I'll add this detail into our existing docs.
Generation of any kind (Brief, sitemap, wireframes, design)
Edits (you can prompt AI to make edits to your site)
Publishing - we scan the site and find issues
Where AI usage doesn't factor in, are any manual edits that you are making. So I think balancing prompting AI to make changes and using the editor we've built, would be key.
Issac A.
·
·
Cool makes sense and thank you and we both understand each other now so it's all that matters. And yeah I don't take it as a personal challenge but I'll probably be one of your pro x20 users eventually that's what I meant about being responsible. I think I've used about 5% and to me it wasn't much considering what was generated which is one page maybe two the brief wireframe sitemap etc. That's not a complaint that's just me observing the usage. Like I said thanks for the heads up just didn't want to guess.
Matt J.
·
·
yeah the initial lift definitely burns up some usage quickly - we're tracking this though so keep us in the loop with what you are seeing and feeling around usage - I'm sure we'll need to make some adjustments
💙1

Matt J.
·
·
Issac A. not entirely sure we'd share costs - I'll ask or someone can step in to answer that directly but I do know that at the end of the day, we're monitoring AI usage and are willing to adjust accordingly. I think the last thing we want to do is block work from happening because of token usage - we just have to be reasonable about eating costs. Its why we deliberately avoided having a token counter or usage meter. AI usage should be a non-issue, something you dont have to think about or worry about. The AI team have massive dashboards that are tracking various metrics and cost is one of them - a small percentage of new users are hitting the limit on the free plan right now, and even less people are hitting the limits on the Pro 5x or Pro 20x plans. In other words, push it and let's adjust. We're listening!
Matt J.
·
·
also should be said that Relume is indeed eating some costs on the free plan with AI - we're willing to do this so that users have a solid opportunity to test drive.
Issac A.
·
·
Uhhh, seems like I might have communicated this poorly again. All I want to do is know what features generally use AI credits, like prompting the brief, or generating a page. I don't want to know the API costs to you, l lol
Matt J.
·
(edited)·
ahhh - hope its not me, not comprehending your question correctly haha. I'll be honest, I try to read the questions a couple of times to make sure but yeah. So yeah, a lot of things touch AI - we're in the tens of billions of tokens used up already across the system. I'll add this detail into our existing docs.
Generation of any kind (Brief, sitemap, wireframes, design)
Edits (you can prompt AI to make edits to your site)
Publishing - we scan the site and find issues
Where AI usage doesn't factor in, are any manual edits that you are making. So I think balancing prompting AI to make changes and using the editor we've built, would be key.
Issac A.
·
·
Cool makes sense and thank you and we both understand each other now so it's all that matters. And yeah I don't take it as a personal challenge but I'll probably be one of your pro x20 users eventually that's what I meant about being responsible. I think I've used about 5% and to me it wasn't much considering what was generated which is one page maybe two the brief wireframe sitemap etc. That's not a complaint that's just me observing the usage. Like I said thanks for the heads up just didn't want to guess.
Matt J.
·
·
yeah the initial lift definitely burns up some usage quickly - we're tracking this though so keep us in the loop with what you are seeing and feeling around usage - I'm sure we'll need to make some adjustments
💙1