DeepMake provides a desktop application that runs open-source generative AI tools locally, enabling visual creators to generate, edit, and upscale images and videos without usage limits or cloud fees. The software integrates directly with Adobe After Effects, offering features like text-to-image generation, smart face and object masking, and AI-powered upscaling. Users can select from various open-source models, including Stable Diffusion and SegmentAnything, through a built-in model manager.
- Artificial Intelligence
- AI Agents
- Creator / Creative Economy
- Media & Entertainment
- Software Only
Funding
Founders
Product
Problem
Visual creators often rely on cloud-based AI tools that impose usage limits, recurring fees, and reduced control over the creative output. This dependency can slow down workflows, increase costs, and restrict experimentation with the latest generative AI capabilities.
Solution
DeepMake brings the power of open-source generative AI directly to the user's computer, eliminating usage caps and cloud dependencies. The software integrates with leading VFX and video editing tools like Adobe After Effects, allowing creators to apply AI-driven effects within their existing workflows. By running locally, DeepMake provides higher control over the output and ensures privacy, while a monthly update cycle introduces new AI tools and models. The platform supports a range of functions including text-to-image and image-to-image generation, smart masking for faces and objects, and AI-powered upscaling, all powered by models like Stable Diffusion, SegmentAnything, Bisenet, and BasicSR.
Target Audience
Primary users are visual effects artists, video editors, and content creators who use professional software like Adobe After Effects and need powerful, unlimited AI tools integrated into their local workflows.
Features
- Local execution of AI models on the user's own GPU, ensuring no usage limits or usage-based fees
- Seamless integration with Adobe After Effects, with Nuke and DaVinci Resolve integrations planned
- Text-to-image and image-to-image generation powered by Stable Diffusion models
- Smart masking for faces using Bisenet and for any object using SegmentAnything and GroundingDINO with text prompts
- AI-powered image and video upscaling by up to 4x using BasicSR SuperResolution models
- Models Manager that allows users to select from vetted open-source models or download and run their own from repositories