Learn to make stickman animations that stay editable.

How to work with AnimCreator: your first animation from a text prompt

AnimCreator is a browser-based stickman animation maker, which means two things in practice: there is nothing to install, and the result of an AI generation is never a locked video file. When you turn text to animation here, you get a timeline full of editable clips — motion, camera, expressions, audio — that you can keep reshaping until the scene reads the way you imagined it. This guide walks through a first animation end to end.

Step 1 — Describe the scene in the prompt box

Open the studio and you land in the Scene panel, with a prompt box front and center. Write what should happen the way you would describe it to a friend: the action, the mood, and any staging you care about. A prompt like "a boxer bounces on his toes, dodges twice, then lands a big counter punch" gives the generator far more to work with than just "boxing" — it produces distinct beats that become separate clips on the timeline, each one editable on its own.

Direction words help too. Mentioning that a character walks in from the left, reacts slowly, or that the camera should zoom in on the landing all translate into real timeline structure: travel direction, speed modifiers, and camera clips.

Step 2 — Pick how much the AI decorates

Next to the prompt sits an AI detail setting with three levels. Lite generates motion, direction, camera moves, and face expressions — the core performance. Assisted also lets the AI pick a fitting background for the scene. Full goes further and dresses the character with wearables and handheld props from the built-in wardrobe. If you already know exactly which background and outfit you want, stay on Lite and set them yourself from the sidebar; if you want a one-prompt scene, go Full and adjust afterwards — everything the AI chooses remains swappable.

Step 3 — Decide whether the character speaks

The voice mode selector controls how audio enters the scene. With voice off, your prompt drives motion only. In Motion + Voice mode you write two texts — one describing the movement, one containing the spoken lines — which is ideal when the action and the words are different things. In Script → Motion mode you write a single script and AnimCreator both voices it and choreographs motion to match, which is the fastest route to a talking, voice-synced animation. Generation costs tokens based on how much text you send (spaces are free), and new accounts start with a balance to experiment with.

Step 4 — Generate, then look at what you actually got

Press Generate and the scene plays. The important habit to build: look down at the timeline editor under the canvas. Each phase of your prompt is a motion clip; camera moves, expressions, and the voice track have their own rows. This is the moment AnimCreator stops being a generator and becomes an animation editor — nothing you see is final.

If the motion is right but the delivery is off (or the other way around), you don't have to pay for a full re-roll. After the first generation, regenerate controls appear that let you redo only the motion while keeping the recorded voice, or re-record only the voice while keeping the motion — you edit the relevant prompt and rerun just that half.

Step 5 — Make three small edits

1. Fix one timing. Drag a motion clip's edge to trim a slow beat. The scene instantly rebuilds with smooth transitions between clips.
2. Add one reaction. Drop an expression clip — a smirk, a shocked face, a glare — over the moment that needs it. Expressions layer over any body motion.
3. Set the stage. Open Backgrounds in the sidebar and pick an environment; a plain white scene and a night rooftop tell very different stories with the same motion.

The timeline editor: trim, retime, and layer your animation

Every serious edit in AnimCreator happens in the timeline editor at the bottom of the studio. It works like the timeline in any online animation editor you may have used — stacked tracks, draggable clips, one playhead — but it's built around generated motion, so the interesting parts are what happens when you stretch, shorten, and swap clips the AI created.

Reading the tracks

• Motion clips are the body performance — each beat of your prompt, or each motion you placed by hand.
• Camera clips hold pans and zooms, so framing is timed like everything else instead of being a global setting.
• Expression clips override the face for their duration and snap back afterwards — reactions live here.
• Audio clips carry voice and sound with a waveform you can slide and trim against the motion.
• Particle and drawing rows time your effects and canvas drawings; every drawing gets its own row, so clips never pile up on top of each other.

Stretching and shortening clips — the four modes

Dragging a motion clip's edge asks a real question: what should fill the extra time, or what should give way? AnimCreator answers with four clip modes, switchable from the popup that opens when you click a clip:

• Loop — an extended clip repeats its motion, with the seam between repeats automatically blended so it doesn't pop.
• Slow — the same motion is spread across the longer duration, great for dramatic moments.
• Speed up — a shortened clip plays its full motion faster, which is how you make a fight snappier.
• Cut — a shortened clip simply ends early, dropping its tail.

The same popup also lets you swap the motion itself — search the library, click a replacement, and the clip keeps its position and timing with new content inside. Between neighboring clips, the editor eases the first frames of each incoming clip from wherever the character actually is, so rearranged sequences don't teleport between poses.

When clips aren't enough: frame editing

Sometimes one pose in the middle of a generated motion is just wrong — an arm too low on a punch, a landing that doesn't quite plant. The Edit frames button on the timeline switches the strip into a filmstrip of every baked frame. Click a frame, then drag the character's joints right on the canvas to fix the pose. When you release, your correction is automatically smoothed into the neighboring frames so the fix doesn't look like a glitch, and it stays inside its clip so the surrounding motion is untouched. It's the closest thing to keyframe animation without leaving the generated result.

A practical polishing pass

1. Play the whole scene once without touching anything, and note where your attention drops.
2. Tighten those moments: shorten the clip and choose speed-up, or cut a redundant beat entirely.
3. Slow down the single most important pose — contrast in pacing is what makes stickman animation feel choreographed rather than generated.
4. Add a camera clip on the key action; a modest zoom does more than any effect.
5. Nudge the audio clip until the voice lands exactly on the gesture. Use the timeline zoom control to get frame-precise; the view follows the playhead automatically so you never lose it while zoomed in.

The sidebar and Tools palette: everything around the canvas

The left edge of the studio is a slim icon rail — the sidebar menu — that switches the panel beside it between five categories. The design goal is simple: the canvas never disappears, so you can restyle a character or swap a background while watching the animation it affects. Here's what lives in each category, and where the on-canvas Tools palette fits in.

Scene — where every animation starts

Scene holds the generation prompt, the AI detail level, voice modes, the entry into Dialogue mode, and the camera and output settings, including aspect ratio. If guide 01 covered your first generation, this is the panel you were in the whole time. It's also where you come back to regenerate half a scene or adjust output before export.

Character — outfits, props, and a supporting cast

Character is the wardrobe and casting department. Built-in skins and items are organized in slots — headwear, face, body, legs, footwear, hands, held item, plus the body color — and the slots stack: a helmet, a chestplate, and a frying pan in hand can all be equipped at once because they occupy different slots. Everything previews live on the canvas.

The item editor goes beyond the built-ins: upload your own image, pick which body part it belongs to, and it attaches to that part of the rig and moves with the animation (there's a full walkthrough of putting your own drawing on a character). Your uploads are saved to your account and show up in the picker under ★ My items. This panel is also where you add extra characters — each with its own motion prompt, voice, outfit, and position — which is what makes duets, fights, and conversations possible.

Backgrounds and Particles — the environment

Backgrounds offers a searchable, filterable library of scenes — interiors, city, nature, weather moods — right inside the sidebar; clicking a swatch applies it instantly, and a paper-color picker tunes the base canvas tone. Particles adds motion effects: dust, sparks, smoke, explosions, and animated sprite sheets. You place a particle on the canvas, position and scale it, and it gets its own timeline clip so the effect fires exactly when the action needs it, facing whichever direction you choose.

Studio — hand-authoring motion and faces

Studio is for creators who want to build motion rather than only generate it. The Create subtab is a keyframe editor: pose the stickman by dragging joints (with an optional joint-angle lock that keeps elbows and knees anatomically sane), step between frames with A and D, and copy-paste whole keyframes along the strip. Save the result under a name and it joins your motion library — from then on, typing that name in a Generate prompt uses your motion. The Faces subtab is the same idea for expressions: a point-based editor where you sculpt eyes, brows, and mouth frame by frame, then save custom faces that appear in the expression picker under My faces.

Tools — drawing as part of the animation

The Tools palette lets you draw directly on the canvas: freehand strokes, shapes, and text labels. What makes this more than a scribble layer is that once a scene is generated, each drawing gets its own clip row on the timeline, so an arrow can appear only during the move it points at, an impact flash can flash for six frames, and a caption can hold exactly as long as it's being read. In explainers the drawings guide attention; in comedy shorts a hand-written label is often the punchline.

Voice, dialogue, and finishing your scene

Silent stickman animation reads as a sketch; a voiced one reads as a short film. AnimCreator generates voice-over from your script and keeps it synced to the motion on the timeline, so adding speech is a writing task, not an audio-engineering one. This guide covers solo narration, directing the voice performance, multi-character dialogue, and what "finished" looks like.

One character speaking

From the Scene panel, the two voice modes from guide 01 apply: Script → Motion turns a single script into both the voice and matching motion — the fastest way to a talking scene — while Motion + Voice keeps the spoken lines and the physical action as separate texts, for when a character says one thing and does another. Either way the generated audio lands as a clip on the timeline, where you can slide and trim it against the motion like everything else. The voice-over deep dive walks the whole pipeline step by step.

Directing the performance with voice tags

A flat read can undermine a great scene, so scripts support performance tags. Inline tags such as [pause] or [sigh] drop a beat or a non-verbal sound exactly where you place them; wrapping tags such as <loud>…</loud> or <whisper>…</whisper> change how a whole phrase is delivered. You don't need to memorize them — every script field has a tag toolbar that inserts them at the cursor, and wrapping tags place the cursor between the opening and closing tag so you just keep typing.

Conversations: Dialogue mode

For two or more characters talking, switch to Dialogue mode. You write the conversation as a list of lines, each assigned to a character, and optionally give each line its own motion prompt — a shrug on the excuse, a step forward on the accusation. AnimCreator voices each line, animates each speaker in turn, keeps the camera able to focus on whoever is talking, and renders subtitles that follow the audio. The result plays as one continuous scene, and it stays editable: each line's audio is a clip, and dialogue has the same half-regeneration controls as the main generator, so you can re-voice one exchange or re-choreograph the motion without rebuilding the whole conversation. For a full worked example, see how to animate a character talking to another character.

Finishing the scene

A reliable final pass looks like this:

1. Watch it cold. Play the scene from frame zero without touching anything. Your ear will catch sync drift faster than your eye.
2. Sync the peaks. Zoom into the timeline and align the loudest word with the strongest gesture — one well-landed sync point makes the whole scene feel deliberate.
3. Check the frame. Confirm the aspect ratio in Scene's output settings matches where the animation is going — widescreen for YouTube, vertical for Shorts and Reels (the short-form guide covers vertical staging).
4. Export. Render the finished scene from the output controls, ready to share.

And because every scene autosaves as a project in My Projects, "finished" is never final: reopen the same project next week, swap the script, and export a new variation of the same staging in minutes.