Plan start and end frames, tune the timing, and copy a transition brief. Free local text tool—not a video generator.
0s · Start frame
Front-facing mug, handle visible on the right, entire base and lid in frame.
4s · Middle frame
The mug has rotated a quarter turn; the same handle is visible in profile.
8s · End frame
The mug completes a half turn; hold a clean rear view with the handle on the left.
Subject and reference notes: An unbranded matte green travel mug on a neutral tabletop. Target duration: 8 s Target aspect ratio: 16:9 Keyframe timeline 0s — Start frame: Front-facing mug, handle visible on the right, entire base and lid in frame. 4s — Middle frame: The mug has rotated a quarter turn; the same handle is visible in profile. 8s — End frame: The mug completes a half turn; hold a clean rear view with the handle on the left. Motion between frames: Rotate the mug steadily around its vertical axis. Slow down smoothly before the end frame; no cuts. Camera path: Locked eye-level close-up with the same focal length. Keep consistent: Same mug shape, green color, lid, handle, scale, tabletop and light direction. Avoid: Extra handles, shifting logos, warped rims, floating base or changing background. Sound cues (optional): Quiet room tone; no speech or music. Planning brief only: attach authorized reference images separately and confirm timing support in the chosen video tool.
No image upload, video rendering or credit charge happens here. Times and aspect ratios are planning targets, not guaranteed model settings. Check supported keyframe inputs in the video tool you use.
Describe what is visible at each moment, not just the mood. Keep the same subject count and framing unless the change is intentional. For a product rotation, a front view and rear view need a believable path between them. Use authorized reference images in the rendering tool; this planner only handles text.
An optional middle frame can anchor a turn, doorway crossing or reveal. Move it to 25%, 50% or 75% of the target duration. The timeline converts that position into seconds, so a 25% waypoint in a 12-second plan appears at 3 seconds. These are editorial targets, not enforced API parameters.
Separate object motion from camera motion. List details that should remain stable, then review the rendered result for changing geometry, flicker, cropped subjects and sudden jumps. If the transition fails, simplify the motion or use closer starting and ending compositions rather than adding more instructions.
No. This independent NanoPic tool builds a text keyframe brief in your browser. It does not call FLUX 3, upload files or render video, and it is not affiliated with Black Forest Labs. Use the official model link for current access. NanoPic's separate video generator uses the model shown in that tool.
Two-frame mode describes the starting and ending states. Three-frame mode adds an intermediate visual anchor. Turning the middle frame off removes its description and timestamp from the copied brief, while keeping your edit available if you turn it back on.
The start stays at zero, the end moves to the selected duration, and the middle frame keeps its relative position. For example, halfway through an 8-second plan is 4 seconds; halfway through a 12-second plan is 6 seconds. Confirm what timing controls your actual rendering tool supports.
No. Describe the reference frames here and attach the actual images separately in a compatible video tool. Use only images you own or have permission to use. A textual frame description is not a substitute for an uploaded keyframe image.
Black Forest Labs' August 4, 2026 release describes text- and image-based video generation, keyframes, continuation and native audio. That is model context, not a statement that NanoPic has connected those capabilities here. Consult the official source for current access and supported inputs.
No. Product turn, Room reveal and Paper bloom are editable planning examples. They demonstrate frame order and continuity notes, not model performance. The copied text cannot guarantee a particular motion, sound, identity match or exact timestamp in a generated video.
