Pixazo launches Agent P, a chat-based AI that picks the model for you
Pixazo has launched Agent P, a conversational AI agent that produces images, video, visual effects and audio from requests written in everyday language. According to the company, it chooses the right model from more than 200 and delivers the finished result from a single conversation.
- Pixazo launched Agent P, a conversational AI agent for images, video, VFX and audio.
- Agent P automatically routes each task to a model from Pixazo's library of 200+ and shows which was used on request.
- It covers micro-dramas and series with consistent casts, plus VFX such as rotoscoping, relighting and crowd generation.
- Agent P is available now on the Pixazo Playground, with no setup required, according to the company.
A conversation instead of menus and settings
Pixazo says Agent P removes the need to select models, write technical prompts or manage multi-step workflows. Users brief the agent the way they would brief a colleague, describing the subject, mood, format or reference. The agent asks clarifying questions when needed, keeps context across the conversation and accepts corrections and refinements mid-session.
The company states that there are no menus, settings or model documentation to work through: "the conversation is the whole interface." Agent P is available now on the Pixazo Playground at playground.pixazo.ai/playground/agentic, and according to Pixazo no setup is required.
What Agent P can make
Pixazo lists several areas the agent covers:
- Images: generating from text, editing and restyling photos, swapping backgrounds, building character-consistent visuals and staging product shots.
- Video: producing clips from text or images, animating stills, stitching multi-scene sequences with consistent characters, and creating short-form content for social and advertising.
- Visual effects: rotoscoping and alpha mattes that hold fine detail such as hair, fur and smoke; background replacement with every layer returned separately; crowd generation and duplication; relighting; and cleanup.
- Micro-dramas and series: taking a brief or script through storyboards, locked characters and locations to finished vertical episodes, with the same cast consistent from episode to episode.
- Audio: adding or replacing a video's soundtrack in the same session.
Automatic model routing and planning
Behind the chat, Agent P routes each task to the best-suited model in Pixazo's library of 200+ models from leading AI providers. The routing is automatic, but the agent shows which model was used for users who want to know.
For complex requests that span several steps or media types, Pixazo says the agent plans the full sequence before starting and reports progress at each stage. Users can redirect, refine or restart at any point without losing context.
Abhinav Girdhar, Founder and CEO of Pixazo, framed the product around a long-standing obstacle: "The gap between having a creative idea and being able to make it has always been the problem. Agent P closes that gap through conversation."
About Pixazo and what it means for users
Pixazo is a generative AI platform that gives individuals and businesses access to 200+ AI models for images, video and audio. For studios and production companies, it also offers AI VFX services that are human-finished and delivered to broadcast spec. These include rotoscoping, compositing and keying, set extension, crowd duplication, face replacement, AI dubbing with lip-sync and previs. Pixazo is a brand of Appy Pie LLP.
For users, the pitch is that choosing among hundreds of models and chaining steps together is handled by the agent. Anyone with a creative brief can start a session on the Playground and describe the result they want. The release does not give pricing or performance figures, so how Agent P compares with working directly with individual models is not stated.
