DRISHNEXA
AI video generation models
The AI writing, voice, image and video generation models used inside the DRISHNEXA studio — and the human approval gates around them.
Writing and screenplay
Writing models that turn a story idea into a Hindi/English scene outline, dialogue and scene plan. Every draft is editable and the writer makes the final call.
google/gemini-3.6-flashopenai/gpt-6-astra
Voice and speech
A text-to-speech model that creates synthetic voice samples for character lines. A real actor's recorded voice always comes from the approved voice bank — a recorded voice is never cloned.
google/gemini-3.1-flash-tts-preview
Image and poster
An image-generation model for character reference sheets, scene visualisation and poster drafts. All drafts are fictional synthetic designs — never imitations of real people.
openai/gpt-image-2.5-sunburst
Video generation
A model that turns scene descriptions into short video clips for the studio's shot planning. Every generation spends AI credits and the output stays private.
google/gemini-omni-1.1-flash
Approved face-and-voice shots
The avatar model (via Replicate) runs only when the character identity is locked, final rights approval is granted and a consent-bearing real recording exists. Without these conditions, rendering stays blocked.
kwaivgi/kling-avatar-v2
Availability depends on service providers, permissions and approvals. Model versions may change over time; this page reflects the current studio configuration.