--- title: "How to generate an image for a post with AI in Growf | Growf" description: "Ask Growf AI for a visual, pick an image model, check the credits per image and click Generate. The image lands in the media library and on the post." url: "https://www.growf.io/resources/guides/generate-an-image-for-a-post/" language: "en" --- # Generate an image for a post [![Wesley Blokhuis – CEO Growf](https://www.growf.io/media/uploads/2026/08/Wesley-Blokhuis-CEO-Growf.jpeg?width=104 "Wesley Blokhuis – CEO Growf")](https://www.growf.io/auteurs/wesley-blokhuis/) [Wesley Blokhuis](https://www.growf.io/auteurs/wesley-blokhuis/) CEO [](https://www.linkedin.com/in/wesley-blokhuis-32642459/) Ask the chat for a visual, pick one of your agency's image models and click Generate. The image goes into the media library and onto the post. The chat can make an image for a post, email or article with one of the image models your agency switched on. It works in Content mode and in the other chats of a workspace. You need permission to write in the chat to ask for an image, and permission to add files to the workspace to generate it; Admin and Strategist have both by default. 3 steps · About 3 minutes 1 ## Ask the chat for an image Open the post you want a visual for, from its project, the calendar or Search projects and items… on Content projects. In the chat box at the bottom, which reads Ask AI to rewrite, or select text…, describe what the image should show and send it. Say which format you need when it matters: square for a feed post, portrait for Instagram or stories, wide for a LinkedIn or blog header. Because the post is open, the chat knows which post the image is for. ![Ask the chat for an image – Generate an image for a post](https://www.growf.io/media/uploads/2026/10/growf-kb-en-generate-an-image-for-a-post-ask-for-an-image.webp) 2 ## Read the image proposal The chat does not make the image straight away. It writes a prompt for the image model and shows it in the card Image ready to generate. The format is at the right of the card's top line, here 1:1\. A long prompt folds to a few lines: Show full prompt opens all of it. The prompt is in English; any text the image has to show stays in the language of your content. You cannot edit the prompt in the card, so tell the chat what should be different and it proposes a new one. When the conversation is folded, Show conversation opens it. ![Read the image proposal – Generate an image for a post](https://www.growf.io/media/uploads/2026/10/growf-kb-en-generate-an-image-for-a-post-review-the-image-proposal.webp) 3 ## Pick a model and click Generate Under the prompt you see the image model the card starts on and its price, here ≈ 43 credits per image. Click the model to pick another of the models your agency switched on; each one shows its own estimate. Then click Generate. The card reads Generating image… and says this can take up to a minute. Next it reads Image generated, shows the picture and the line Added to the media library as, followed by the file name; click the picture to open it in a new tab. The chat then adds the image to the post, next to any media it already had, and its card reads Media set on 1 item; you find the image in the Media section of the post. ![Pick a model and click Generate – Generate an image for a post](https://www.growf.io/media/uploads/2026/10/growf-kb-en-generate-an-image-for-a-post-pick-a-model-and-generate.webp) ## FAQ Why does the chat say No image models switched on?+ Your agency has no image model switched on, so the chat cannot make images. The agency owner or an agency admin switches them on in Agency cockpit > AI models, on the Image models tab, and chooses the model under Images start with. Then ask the chat for the image again. Why can't I make an image in a personal chat?+ Images are added to a workspace's media library, so the chat only makes them in a chat of that workspace. In a personal chat the card reads Not available in a personal chat. Ask for the image in a chat of the workspace it is for, for example in Content mode with the post open. What if the image is not generated?+ The card says what went wrong, and Generate turns into Try again. When the model returned nothing, for example because it declined the prompt, the card reads The model returned no image, often because it declined the prompt. Pick another model before you click Try again, or ask the chat for a different image. When a model is no longer available to you, pick another one. Each card makes one image; for another version, ask the chat again. What does an image cost?+ Each image uses your agency's AI credits, and a workspace can have its own credit limit. The card shows an estimate per image, and every model in the list shows its own; what you pay follows what the image actually cost. Asking the chat for the prompt and adding the image to the post use credits too, like any chat message. When the credits are used up, the card says so: Your agency has used its AI credits, or This workspace has used its AI credits. An agency admin sees where the credits go in \*\*Monitor agency AI usage\*\*. Why can't I generate the image or pick a model?+ Asking the chat needs permission to write in the chat. Generate also needs permission to add files to the workspace: Admin and Strategist have it by default, Creator does not and sees You do not have permission to add files to this workspace. When your role may not pick models itself, set under Access per role in Agency cockpit > AI models, the card names the agency's default image model and shows no list.