I haven't used ChatGPT yet because I started with Claude, which has worked very well for me so far. Based on this thread it seems that ChatGPT is better at image generation. What type of account do I need for ChatGPT? I assume the free account has limitations, and that I would need a Pro account for image generation.
You don't need a ChatGPT subscription for this. What you need is an OpenAI API key with some credits on your OpenAI account. Once that's set up, you can generate ChatGPT Images inside Claude through Amplifiers.
Hi! I set up the MCP and tried to use the ChatGPT image integration for the first time but got this... "Contact whoever runs the "AI Blew My Mind" plugin (the tool powering image generation here) and ask them to top up the OpenAI API credits on their end. That's the account that's actually running dry."
Hey Allison! I cover the first 3 generations for each user out of my own pocket so people can test it properly first. If you want to continue using GPT Images 2 after that, you’ll need to add your own OpenAI API key at:
Appreciate this a lot, one of the best compliments you could give me. Thank you 💜
It’s really been super useful for me too. Just last night we added image generation with ChatGPT Images 2 as well (which is now the best model).
And for my work, I’ve set up skills for things like thumbnails and infographics for my articles, and now it’s just… frictionless. I don’t have to think about prompts, I don’t have to switch tools, and the output comes out right from the first try because of how the prompts are embedded behind the tool.
It’s saving me a ton of time on my side too, so I’m really glad it’s useful for you as well.
I use Flow from Google, which I guess is Gemini/Nano Banana Pro 2, for my illustrations on Millennial Masters. I can download them upscaled at 2k, which is better than ImageFX (also Google's tool that's being retired and I was using before). What I like about Flow is that you can animate them, and hopefully soon we can have more moving pictures across Substack.
GPT is certainly a no-go. There's something about those visuals that I can tell from a mile that they've been made with GPT, and they're all very samey. Although... some people can have very good prompts and turn them into some amazing and consistent visuals.
From what I've seen with you and a couple of others, Claude it's pretty good at it now too, especially these infographics. But I can't get my head around how these infographics really work and are useful, considering most people consume on mobile.
Flow is great, I actually thought it was mainly for video, that’s what I used for, didn’t realize you can use it for images too. I also used Google AI Studio instead of Gemini mostly for the resolution. GPT has a new model out, might have gotten better.
Claude actually doesn’t have native image generation, we added that through Amplifiers, and it’s also using Nano Banana 2 behind the scenes.
I use infographics a lot for my articles and LinkedIn. And we added prompt generators behind Nano Banana in Claude, so you can generate all kinds of images without having hustling with the prompt.
Thanks for sharing. Everything connected well as per your excellent instructions. Am on the Pro plan. But on my first try, having just reset my usage limits, Claude said 'no'.
"Hit a quota wall on the image generator — the free tier daily limit for that model is exhausted."
Awesome, glad you're in! That error is your Claude usage quota, nothing to do with Amplifiers.
If you're on the Pro plan, the limits are more restrictive on how much you can use certain models. Just try switching to a different model when you run into this. You were probably on Opus. Click on the model name in the bottom right of your chat window, and pick Sonnet from the model picker instead. That should get you going.
i've been wanting to try it but claude keeps giving me this, despite me having set up the connector properly: There's an "Infographic generator" amplifier in the catalog, but it's on the Pro tier and your account doesn't have access to it. The ones you do have access to (LinkedIn Post Generator, Content Repurposer, etc.) aren't built for infographic image generation.
Hey Magali, thanks for letting me know! I just checked the email address you used for your AIBMM subscription, and there's no account using it on auth.aiblewmymind.com. So if you created your auth account with a different email address, that's why Claude doesn't see your premium access.
What you'd need to do: create a new account on auth.aiblewmymind.com using the same email address as your subscription. Then disconnect the MCP in Claude and reconnect it with the new account. The premium access is tied to Stripe, so it needs the email addresses to match in order to know you're a premium subscriber.
I haven't used ChatGPT yet because I started with Claude, which has worked very well for me so far. Based on this thread it seems that ChatGPT is better at image generation. What type of account do I need for ChatGPT? I assume the free account has limitations, and that I would need a Pro account for image generation.
Thanks
You don't need a ChatGPT subscription for this. What you need is an OpenAI API key with some credits on your OpenAI account. Once that's set up, you can generate ChatGPT Images inside Claude through Amplifiers.
To set it up:
1. Go to http://platform.openai.com and sign in.
2. Create an API key under API Keys.
3. Add payment details under Billing and purchase some credits (for example $10).
4. Add the API key at auth.aiblewmymind.com.
One nice thing is that with the API you're paying for what you use. With $10 you can generate around 200 images.
I also put together a step-by-step guide with screenshots if it helps:
https://scribehow.com/viewer/Setting_Up_an_OpenAI_API_Account_and_Project__mbjacQIjQaeeOK_lGMHxiw
I appreciate the response, and the guide was excellent.
Thanks.
Hi! I set up the MCP and tried to use the ChatGPT image integration for the first time but got this... "Contact whoever runs the "AI Blew My Mind" plugin (the tool powering image generation here) and ask them to top up the OpenAI API credits on their end. That's the account that's actually running dry."
Hey Allison! I cover the first 3 generations for each user out of my own pocket so people can test it properly first. If you want to continue using GPT Images 2 after that, you’ll need to add your own OpenAI API key at:
https://auth.aiblewmymind.com/
That way the image generations run on your own API credits. The cost is relatively low, roughly around $1 per ~20 images.
Hi! I did that but got this error on my very first try 🫠
ugh, could you maybe try again and DM me a screenshot or send me an email at daria@aiblewmymind.com? will look into it so we can sort it out together
Emailed you! Thank you :)
Wow! What I love about your stuff, Daria, is you don't do anything just to do something cool...this literally saves tons of time.
Appreciate this a lot, one of the best compliments you could give me. Thank you 💜
It’s really been super useful for me too. Just last night we added image generation with ChatGPT Images 2 as well (which is now the best model).
And for my work, I’ve set up skills for things like thumbnails and infographics for my articles, and now it’s just… frictionless. I don’t have to think about prompts, I don’t have to switch tools, and the output comes out right from the first try because of how the prompts are embedded behind the tool.
It’s saving me a ton of time on my side too, so I’m really glad it’s useful for you as well.
I use Flow from Google, which I guess is Gemini/Nano Banana Pro 2, for my illustrations on Millennial Masters. I can download them upscaled at 2k, which is better than ImageFX (also Google's tool that's being retired and I was using before). What I like about Flow is that you can animate them, and hopefully soon we can have more moving pictures across Substack.
GPT is certainly a no-go. There's something about those visuals that I can tell from a mile that they've been made with GPT, and they're all very samey. Although... some people can have very good prompts and turn them into some amazing and consistent visuals.
From what I've seen with you and a couple of others, Claude it's pretty good at it now too, especially these infographics. But I can't get my head around how these infographics really work and are useful, considering most people consume on mobile.
Flow is great, I actually thought it was mainly for video, that’s what I used for, didn’t realize you can use it for images too. I also used Google AI Studio instead of Gemini mostly for the resolution. GPT has a new model out, might have gotten better.
Claude actually doesn’t have native image generation, we added that through Amplifiers, and it’s also using Nano Banana 2 behind the scenes.
I use infographics a lot for my articles and LinkedIn. And we added prompt generators behind Nano Banana in Claude, so you can generate all kinds of images without having hustling with the prompt.
Oh, so that’s why everyone was excited about Claude doing native images this week.
Makes sense, and indeed, Flow is good with Nano Banana 2. Tried the same prompt (written by GPT) in v1 and it was ridiculous.
Thanks so much. This is so helpful. I just generated my first image with Amplifiers and was very pleased.
I’m so glad you liked it, Randy!
Thanks for sharing. Everything connected well as per your excellent instructions. Am on the Pro plan. But on my first try, having just reset my usage limits, Claude said 'no'.
"Hit a quota wall on the image generator — the free tier daily limit for that model is exhausted."
??
Awesome, glad you're in! That error is your Claude usage quota, nothing to do with Amplifiers.
If you're on the Pro plan, the limits are more restrictive on how much you can use certain models. Just try switching to a different model when you run into this. You were probably on Opus. Click on the model name in the bottom right of your chat window, and pick Sonnet from the model picker instead. That should get you going.
Let me know if it works!
Guess what I am trying today after I create my latest article.
yay! please do tell me how you liked it. I know you create awesome visuals for your articles, so I’m super curious
Daria 🤩😊 you’re raising the bar!!
haha, can’t wait for you to try it!
i've been wanting to try it but claude keeps giving me this, despite me having set up the connector properly: There's an "Infographic generator" amplifier in the catalog, but it's on the Pro tier and your account doesn't have access to it. The ones you do have access to (LinkedIn Post Generator, Content Repurposer, etc.) aren't built for infographic image generation.
Hey Magali, thanks for letting me know! I just checked the email address you used for your AIBMM subscription, and there's no account using it on auth.aiblewmymind.com. So if you created your auth account with a different email address, that's why Claude doesn't see your premium access.
What you'd need to do: create a new account on auth.aiblewmymind.com using the same email address as your subscription. Then disconnect the MCP in Claude and reconnect it with the new account. The premium access is tied to Stripe, so it needs the email addresses to match in order to know you're a premium subscriber.
Let me know if it works!