
Before generation you can ask...
For a product ad, the material is often already in meeting notes and brand guidelines. Read that into the context, let the agent mark packaging and copy that must stay, then use brainstorming to shape the script and visual direction. Only after approval do you edit the selected node on the canvas. For publishing or higher spend, switch to ask mode and wait for a yes. Feedback should name both the change and what stays, such as a new background with the bottle text untouched. It fits ad and short-film teams who already have a brief and references and need to deliver on one canvas.
TapNow splits creative work into three parts: an agent that reads the goal and existing material, apps that generate, retouch, edit, or pull in external documents, and a canvas that keeps briefs, references, versions, and final files. Material can be an upload or a read from Lark or Notion. Before generation you can ask for analysis, a script, or a storyboard, and continue only after approval. The docs do not name a developer. The homepage offers a free start and a separate enterprise pricing entry.
Difficulty: Intermediate
Agent breaks down the task
It reads the text, attachments, and canvas nodes you name, confirms the deliverable and constraints, then calls the right app to generate or organize.
Canvas nodes and links
Text, images, video, audio, and local uploads become nodes, and links show which inputs a result used.
Creative and source apps
Official apps generate, retouch, and edit; Lark or Notion apps can read documents so you paste less.
Brainstorm a direction
Brainstorm develops scripts, characters, worlds, and visual directions, and approved notes move onto the canvas.
Pause before generation
You can ask for analysis, a script, or a storyboard only, and use ask mode to wait before spend or publishing.
Ad teams taking a client brief
A team reads Lark notes and brand guidelines, marks details that must stay, then makes frames on the canvas.
Creators revising one node
Select the problem node, change only the cap, name the reference image, and keep light and aspect ratio.
Teams handing off a cut
The agent organizes versions and handoff notes, then the canvas link or final files go back to the workspace.
The agent splits the task from the material and limits you give it, calls apps, and puts results back on the canvas. You can also add text, image, video, or audio nodes yourself, or upload a file. To compare settings, the docs suggest creating a generation node manually. Both kinds of results stay on the same canvas and can be cited later.
Name the one issue for this pass, list what must stay, and point to a reference node. The docs use the example of fixing only the cap when placement is already right, leaving background, light, and aspect ratio alone, and generating one candidate. Approve that candidate before the next pass. For publishing or higher spend, use ask mode instead of generating immediately.
Real reviews and feedback from users