NOTES FROM THE STUDIO

A little direction. A lot of character.

Practical starting points for a distinctive look, a clear scene and a better first video.

What this studio supports

Character-image inputs: Crazy Frog AI uses one photo of yourself, an explicitly consenting adult, a pet or an authorized character, plus the built-in chase video. Select the matching subject mode. Jean Phil supports a single photo or no-photo creation with its selected reference; Hotel Lobby keeps its one-to-two-photo workflow. Use JPG, PNG or WebP within the upload limit shown in each tool.
Video settings: Crazy Frog AI requests 14 seconds at 576P / 768P, in 9:16 / 16:9 / 1:1. Its reference is 16:9, so other formats can change the framing. The inherited Jean Phil studio offers 4–15 seconds. Other tools keep their own model-specific settings.
Preview evidence: The homepage uses a 14-second police-chase motion reference and four scene directions. The inherited Jean Phil gallery has 7 separate video references. Preview clips are attributed third-party movement references, not outputs or customer testimonials from this site. The actual model input is the same pinned reference shown in the tool.
Access and availability: You can view public templates and guides without signing in. Generation requires a signed-in account, enough credits and an available service. Review the quote before submitting; check the existing task if a request is delayed or uncertain. No free-generation allowance or fixed completion time is promised.
Result limits: Generated faces can be stylized and movement or audio can differ from the reference. Exact likeness, choreography, lip sync and high-fidelity face replacement are not guaranteed. Review faces, hands, motion and audio before publishing.
Affiliation and music: This independent tool is not affiliated with Crazy Frog, Jean Philanthrope, reference creators or social platforms. Credits do not include music or third-party character rights. Use media with the necessary rights and explicit consent of adults depicted, and label synthetic output. Unauthorized likeness use and deceptive impersonation are prohibited.
Failures and cash refunds: Confirmed failed generation tasks return the deducted credits. A technically completed result that you dislike is not a failed task. A credit return is separate from a cash refund; purchase refund requests follow the terms and applicable consumer rights.

Video credits by duration and resolution

The anatomy of a Jean Phil moment

Your first character-to-video scene

Write prompts with a clear point of view

FIELD NOTES / 01

The anatomy of a Jean Phil moment

The format works through contrast: an unmistakable character, a completely serious expression, and a scene that does not need to be extraordinary.

Start with a recognisable silhouette

A blunt bob, a curled moustache and a tailored suit create the visual shorthand. Keep the same core description across scenes. You can invent a different colour or costume to make your own recurring character.

Let the setting stay ordinary

A café table, a quiet street or an empty apartment gives the character room to be funny. Describe one location clearly instead of combining multiple places in a short clip.

Pick one readable action

A small dance, two playful jabs or a slow walk works well as a starting point. A short sequence with a clear beginning and end is easier to communicate than a long list of events.

Make my video

FIELD NOTES / 02

Your first character-to-video scene

Start with your own portrait, a photo of a consenting adult, or an authorized fictional-character image.

Choose a clear single portrait

Use a clear photo of an adult who has consented to this use, or an authorized character image. Avoid group images, heavy filters or an obscured face. Obtain the necessary image rights and explicit consent before using another person’s portrait.

Check the frame and the cost

Choose Character image and upload one reference. Select a template video to guide the performance; no ending image is required. Duration follows the template, rounded to the nearest whole second. You can adjust it from 4 to 15 seconds and select 576P or 768P. Review credits before generating.

Review before sharing

Look closely at faces, hands and motion in the completed video. AI output can vary. Download the result you like, label it as AI-generated, and add only music you are permitted to use.

Make my video

FIELD NOTES / 03

Write prompts with a clear point of view

A useful video prompt gives the model a character, a location, an action and a camera instruction. More words are not always more direction.

Character + place + action

Try: An original fictional gentleman in a brown houndstooth suit performs two playful jabs on a sunlit café terrace. His face stays completely serious. Keep the character consistent throughout a single continuous shot.

Describe the camera simply

Choose one instruction, such as a steady medium shot or a slow push toward the character. Avoid asking for multiple cuts, rapid location changes and complex choreography in a very short video.

Iterate one decision at a time

Keep the parts you like and change just the setting, movement or camera direction. Each generation is a new paid request when the service is connected, so check the displayed credit cost before trying again.

Make my video