Adobe Firefly on X

What gets posted most is the output itself: photos put through Firefly come out as paintings, prompt-only sticker sets, PVs assembled from generated images and video in Photoshop, illustrations credited to Firefly's official model. A solo developer describes tuning Firefly alongside other tools to produce 170+ animated card loops on a shoestring budget, and its agent picks image over video from a prompt that does not say which.

The tools draw their own praise: Boards for keeping a project's stages in one place, which one user calls keeping "the pipeline in one place"; custom models trained on a user's own images; Text to vector output downloaded as SVG to edit in Illustrator; sound-effect generation; and video editing with screen and webcam recording and automatic reframing for vertical formats.

5 people point to licensed training data and Content Credentials as reasons to use it for paid work. One creator, uneasy about commercial rights, says the licensed-data explanation made them willing to generate more, and one post relays a report of a creator whose per-image price rose from 50 to 3,000 yen after switching to Firefly for copyright-wary clients. 3 are let down: a beta assistant that will not hold a consistent scene or finish a short film, and quality rated below a rival model.

Output complaints are specific: a jersey whose name came out as "GABRETT", a request for a painting-styled version of a real photo that returned a different person entirely, a transparent background that came back filled with pixels, and a flat refusal to generate the Eiffel Tower.

The unease is about AI in creative work rather than a feature: AI-assisted art sold in direct messages as commissions at over 200 each, AI spreading through Adobe's products, AI reaching performances. Prompting complaints are about not being able to describe the image in mind; one says describing it in words is really hard and the results keep missing.

On price, one sees no sense in paying 22 euros a month and mentions a friend's trouble cancelling, and another objects that a simple drawing costs the same as a complex one — paying for the canvas, not the content. Rough edges are a failure screen one user dislikes and music generation that makes you fill in three separate fields.

Nobody in this section is weighing a single alternative against Firefly; the five themes are level, with one person each, and each makes a different point.

Two go against Firefly on output: one rates Gemini's Nano Banana model above Firefly's own generation but still uses it inside Firefly, and one notes that Firefly will not generate photos of the Eiffel Tower while ChatGPT "seems to just look up the first image".

Two go in Firefly's favour. On upscaling, an image passed through free online tools came out best in Firefly, since "other AI tools hallucinate too much in the details". On price, someone who sells pay-per-image access and finds that cheaper in general makes an exception for Firefly and one other tool, citing "super generous usage". A fifth names Canva and Firefly together as tools that handle image editing better than what someone was considering.

The section holds one post: a short question in Turkish asking whether Adobe Firefly is free ("adobe firefly ücretsiz mi"), put to another account in a reply. It gets no answer in the posts filed here.

With a single question, there is no pattern to describe — no repeated asking about trials, credits, cancellation or which features sit behind a paid plan. A maker reading this should treat it as the shape of the gap rather than a theme: people arriving at Firefly want to know, plainly, what costs money.

The music wish comes from someone using Firefly's song generation who finds filling in three separate fields tedious and wants to work from one prompt, as they already can with sound effects — the same friction another user names as a rough edge.

The other wish follows a run of background generations that kept missing what was in mind: describing it in words is really hard, so they would rather have their thinking turned straight into an image — a future, they add, that feels possible. Both wishes are about getting from an intention to an image with less wording in between; neither is repeated by a second person in the posts filed here.