Storyboard in words: turning a video idea into a frame-by-frame text plan
The 'Storyboard in words' tool converts your video idea into a textual description of frames: what is visible in the frame, how the camera is positioned, what the characters do, and how one frame transitions to another. This is not a video generator and not a photorealistic image generator — the output is structured text that can be used during shooting or passed on to a team. In this article we will explore how to work with the tool iteratively: from the first result to ready-to-use material. You will learn which fields to fill in, how to formulate a request, how to check and improve the draft and what limitations are important to keep in mind. All examples are illustrative and do not represent real shoots.
Open tool: Storyboard planner ↗What exactly does the tool do and which fields to fill in
'Storyboard in words' is a text planner. You describe a video idea, and the tool returns a list of frames with angles, actions and transitions. The interface has three fields: 'Your text or idea' (up to 8000 characters), 'Role or characters' (up to 1000 characters) and 'Length or duration'. Additionally you can choose the output language. There are no sliders, style presets or level of change settings here — just text. It is important to understand: the tool does not draw pictures and does not edit video, it creates a verbal plan.
Fill in the fields specifically. In the idea field describe what is happening: location, object, key action. In the role field specify who is in the frame: one person, a group, an animal or an object. In the duration field set a guideline: '30 seconds', '1 minute', 'a series of three short videos'. The more precise the input, the more useful the draft. For example, for the idea 'a clay mug from a lump of clay to the finished product' it is best to immediately indicate that this is a process, not a plot with dialogues.
After pressing 'Generate' you will get one result in a separate panel. It can be copied or downloaded as a plain text TXT file. The tool also creates vectors (SVG and PNG), speech via eSpeak (WAV) and subtitles (SRT), but these are separate functions not related to storyboarding. Do not expect DOCX or PDF export, batch processing or automatic saving — they are not available.
Example of filling in: idea — 'A short video about how a master molds a clay mug'; role — 'Master, hands close-up'; duration — '45 seconds'.
First result: what you will get and how to read it
The first answer from the tool is a draft. It usually contains a sequence of frames: 'Frame 1: general view of the worker's desk, the master prepares the clay. Frame 2: medium shot, the master forms the walls of the mug. Frame 3: close-up, fingers smooth the edge. Frame 4: finished product on the shelf'. Such text is already useful, but it may miss important details: movement direction, camera position, duration of each frame, sound or dialogue.
Read the result as a basis, not as a final script. Check if it matches your idea. If you asked to show a process but the tool added extra scenes with other characters, that is a signal to refine the input. If the frames are too general ('the master works'), add more specificity to the original text: which exact actions, which body parts or objects are in focus.
Pay attention to transitions. The tool may suggest 'smooth transition' or 'sharp cut', but it does not always explain how to shoot it. This is normal for a text planner: your task is to refine the details to your capabilities. Do not consider the first result final; it is created to save time on structure.
Fragment of the first result: 'Frame 2: medium shot, the master forms the mug walls. Transition: inflow.'
How to refine the request: iterations from draft to plan
The best way to refine the storyboard is to repeat the request with refinements. Do not change everything at once: add one or two details and press 'Generate' again. For example, if the first result is missing angles, write: 'Add for each frame the type of angle: wide, medium, close-up, detail'. If actions are missing, specify: 'Describe exactly what the master's hands do in each frame'. If transitions are missing, ask: 'For each transition specify the method: cut, inflow, fade'.
Keep an eye on the limits. The idea field — up to 8000 characters, the role field — up to 1000. This is enough for a detailed description but not for a full script with dialogues. If your idea is large, break it into parts: first storyboard the first scene, then the second. The tool does not automatically save history, so copy successful variants into a separate file.
Do not ask the tool to do what it cannot do: generate video, photorealistic images, speech with voice cloning or publish material. It works with text only. If the response contains promises 'generate video', this is a model error; ignore such fragments and rely on the textual frame description.
Refinement: 'For each frame add camera angle (top, side, frontal) and duration in seconds'. Result: 'Frame 3: close-up, camera from the side, 5 seconds, fingers smooth the edge.'
Checking the result: checklist before shooting
Before using the storyboard, check it against a checklist. First: are all key moments of the idea reflected? If you wanted to show clay turning into a mug, there should be frames with the raw material, the process and the finished product. Second: are the angles understandable? 'Close-up' without specifying what is in the frame can confuse the operator. Third: are actions described? 'The master works' is too vague; 'the master squeezes clay with palms' is specific.
Fourth: are there transitions? If the plan only has frames but no description of how they are connected, add this manually or refine the request. Fifth: does the duration match? The tool may suggest 10 frames for 30 seconds — that is about 3 seconds per frame, which is realistic. But if there are 20 frames for 15 seconds, the plan may be too granular. Sixth: are there invented details? The model may add something that was not in your idea. Remove the unnecessary.
Remember that model answers may contain errors. The tool does not guarantee correctness and does not fact-check. Your task is to critically read the text and adapt it to real shooting conditions: available equipment, light, location. If something does not fit, change the wording and generate again.
Checklist: 1) is there a frame with the finished product? 2) is an angle specified for each frame? 3) are hand actions described? 4) are there transitions? 5) does the number of frames fit the duration?
Tool limitations and typical errors
'Storyboard in words' is a text assistant, not a studio. It does not generate video and audio, does not create photorealistic images. SVG and PNG vectors — these are vector compositions, not photographs. eSpeak speech — synthesized speech, not a recording of a live voice. SRT subtitles — a separate format, not related to storyboarding. Do not expect DOCX or PDF export, batch processing, automatic saving, integrations or publishing. The interface is a simple web form with text fields.
A typical error — too general input. If you write 'video about nature', the result will be just as general. The more specific the idea, the more useful the draft. Another error — expecting the tool to figure it out for you. It does not know your conditions: how many cameras you have, whether you have a tripod, what the light is like. All of this needs to be added to the request or corrected manually.
A third error — ignoring limits. The idea field is limited to 8000 characters, the role field to 1000. If the text is longer, it may be truncated. A fourth error — using the result without checking. The model may confuse the order of frames or add unnecessary transitions. Always read the response and adapt it to yourself. Queue and tool availability may change, so do not rely on an instant answer in a critical moment.
Error: 'Make a video about a mug'. Better: 'Describe 5 frames: 1) a lump of clay on the table, 2) hands kneading the clay, 3) forming the walls, 4) smoothing the edge, 5) finished mug on the shelf. Angles: wide, medium, close-up, detail, wide. Duration: 40 seconds'.
Practical example: from idea to ready storyboard
Consider a fictional example. Idea: 'A short video about how a master molds a clay mug'. Role: 'Master, hands close-up'. Duration: '45 seconds'. The first request gives a general plan: 'Frame 1: master takes clay. Frame 2: molds the mug. Frame 3: finished product'. This is not enough. We refine: 'Add angles and actions for each frame, indicate transitions'. Second result: 'Frame 1: wide shot, the master sits at the table, takes clay. Transition: cut. Frame 2: medium shot, hands knead the clay, giving it a ball shape. Transition: inflow. Frame 3: close-up, fingers form the mug walls. Transition: cut. Frame 4: detail, the big finger smooths the edge. Transition: fade. Frame 5: wide shot, the finished mug stands on the shelf'. This is already a working plan.
Now we check: do 5 frames fit into 45 seconds? About 9 seconds per frame is acceptable. Are all actions understandable? Yes. Are there transitions? Yes. Does the result match the original idea? Yes. If needed you can add sound: 'background music, sound of molding' — but the tool does not create audio, only text. Speech can be made separately via eSpeak if needed.
This example shows how iterations turn a general draft into a specific shooting plan. Do not be afraid to make several requests. Each time refine one detail: angle, action, transition, duration. Save successful variants in a text file. This way you will get material that can be used immediately on set.
Final storyboard: 'Frame 1: wide shot, the master sits at the table, takes clay. Transition: cut. Frame 2: medium shot, hands knead the clay. Transition: inflow. Frame 3: close-up, fingers form the walls. Transition: cut. Frame 4: detail, smoothing the edge. Transition: fade. Frame 5: wide shot, the finished mug on the shelf'.
'Storyboard in words' is a tool for quickly creating a textual video plan. It does not replace a director and does not generate media, but helps structure an idea. Work iteratively: fill in the fields specifically, get a first draft, refine the request, check the result against a checklist and adapt it to your conditions. Remember the limitations: text only, character limits, possible model errors. A example with a clay mug is fictional, but it shows the logic of refinement. Use the tool as a draft, and always make the final decision yourself.
Open tool: Storyboard planner ↗Interface translations are machine-assisted and may need refinement.