For some reason even changing the text on a button is proving to be an issue 
I am prompting to do things which are 90% less complicated than has been done previously, but this week it just seems like it wants to sabotage 6 months worth of hard work… is anyone else experiencing similar behaviour from v0 this week ?
3 Likes
There haven’t been any widespread reported issues with v0 this week that I’m aware of, but I do have a few troubleshooting suggestions:
- Try being extremely specific in your prompts (e.g., “Change the text of the Submit button in components/Button.tsx from ‘Submit’ to ‘Send’”)
- Use the code editing feature for small changes like button text
- Try breaking complex requests into smaller, more focused prompts
- Refresh the page or start a new chat if v0 seems “stuck” in a pattern
If you’re still experiencing issues after trying those ideas, it would be helpful to see a specific example of a prompt that isn’t working as expected
I have built a fantastic react app using v0 for the last 6 months so I am quite used to it … I have spent around 50 prompts today all of which have resulted in pointless outcomes, even when reduced to simply asking to change the text on a button., I am feeding v0 very granular specifics and the “Thinking” process articulates exactly what I have asked for, however when it rebuilds the page it misses out substantial chunks. For example I created a search user button .. when I bind that button to a function such as (when pressed get users", i produces a page with the button and nothing else … this has happened several times today and it is holding up development .. please help, I love this product and would rather not be disappointed and forced to seek another platform.. I am reluctant to post my code on here but will gladly send it in an email or something
Above is the original before the prompt
The Above is after the prompt (where are the filters? the layout, the design where is my hard work ?) - 50 times I have tried this .. it’s the same thing I have done in about 30 other places in the app with far more complexity ???
Here was the prompt
PLEASE NOTE I DO NOT WANT TO CHNAGE THE DESIGN … simply implement a function that gets the following data and displays it when the current user clicks “Find Professional Services” button then please fetch data foe all users in this table and display them ..
That was the prompt minus the table definition (* for obvious reasons ) … I have explained the tables joins and everything else , which obviously it understood as it’s created something which fetched data from these table , just not what I asked for … If this is going to happen with the new pricing model as if so this is appalling … I have no control over what it does even when I instruct it “NOT TO CHANGE THE DESIGN” I mean how much clearer can I be ?"
I am seeing a lot of other complaints about user paying for these erroneous outputs
I also add this at the end of every prompt “-> this is important please do not change any other parts of the sites look and feel or functionality just focus on that section as it like what you have done . .thank you please do not change DO NOT CHANGE ANYTHING OTHER THAN WHAT I HAVE ASKED PLEASE”
Thanks for sharing these screenshots and details. I sent a DM to get the chat ID from you
1 Like
This was like the 50th prompt so please excuse the caps 
1 Like
No worries! I completely understand your frustration. It’s a really weird situation.
I’m digging into the logs and asking the v0 team for help
1 Like
Same issues for me yday. Got better as day wore on and weirdly was more proficient with backend type work.
1 Like
I still hope to be able to display this proudly and direct more traffic to the platform, one can only hope 
@amyegan we have been having the same issue. No matter what prompt i give, it deletes the whole page and creates a new page again. I have tried this for many pages. This is happening since Sunday.
Can you let me know which model you’re using? Trying to figure out if there’s a pattern here with a specific model.
I have tried both @amyegan and the results remained the same in all instances 
also how are those of us who are not yet on the new plan, intended to be compensated for the down time ? Not meaning to add to what is already on your plate, but just wondered if it’s the intention of the company at some point to address this
1 Like
Wish i could be put back to old plan whilst they iron out the issues and put in some transpency, id like to atleast an estimate of cost per response abd a chance to tailor my instructions to ensure a lesser hit on token cost
we just needed transpency of whats going on and you shouldve weened this process in, we couldved helped mitigate the cost and helped iron out the bugs , you cant just full whack a new pricing model like this, there isnt any safeguarding for wasted tokens, no estimated costs, no eqdy to see per reep9nse value (in chat)
just the basics , youve made such an oversight ! its impossible to miss…
Someone in the vercel department needs firing because this process has been extremely damaging
on reflection @amyegan why am I worrying about this option, when surely it should not yet be applicable to me
I checked with the team, and they’re still investigating. So I don’t have a solution just yet.
@lodos3 Usage depends on the length of your prompt + conversation history and the length of the generated response. You can get a complete list of your message costs at v0.dev/chat/settings/billing
@swatchwatchipd If you’re looking for a refund, please open a case at vercel.com/help so the support team can help 
1 Like
Can you clarify the difference between the two as im trying to understand the rapid decline of my $30
@lodos3 Based on that screenshot:
-
The older message used the original model. I can’t be sure of much else about the message based on the screenshot alone. But it was likely something fairly large based on the cost shown.
-
The more recent message used the new large model, which uses more resources to handle difficult problems and designs. Again, I can’t determine much about the message based solely on this screenshot. But I would guess that you didn’t run the exact same prompt twice three days apart, so we’re not likely looking at an apples to apples comparison between the two.
I hope that helps!
3 Likes
@amyegan let me see how we get on today as I am seeing some interesting result as we speak, looks like our cries did not fall upon deaf ears (not trying to condescend) …
2 Likes
@amyegan no matter which model I try, it is just deleting the code of the file and replacing with the new code. Even to change a word, it would delete the whole page. The platform is basically unusable now.
1 Like
Id like to know if there are plans to reimburse members for empty chats, failed responses and in general v0 errorrs
i can not jusitfy part of my 30 dolllars a month going towards no actual code.
1 Like