Local AI for Photo and Video: The Future of ObviousIdea
At ObviousIdea, we are exploring how local AI for photo and video can help us build useful new products and more intuitive interfaces. Our experiments are already shaping the way we work—and we expect to share more as ideas become features ready for our users.
More than 3 billion tokens processed locally
Between November 2 and 15, 2026, our local text models processed approximately 3.39 billion tokens across input and output. This includes cached input: it measures processing volume, not 3.39 billion newly generated tokens.
We use local AI for software development, research, SEO and search visibility, strategy, and tools that help us monitor our systems. We are also testing AI photo editing and local video generation. The aim is to learn what makes these technologies useful in real products, not just impressive in a demo.
Why explore on-device AI for photos and video?
Local processing can give people more control over where their photos, videos, and other data are handled. It can also make it easier to experiment and iterate. Privacy and security are not automatic, however: they depend on how each application is configured, connected, maintained, and used.
Creative work is especially interesting to us. Editing an image or generating a video often means trying several versions before finding the right one. Local processing gives us a way to explore those workflows and their trade-offs, including hardware requirements and energy use.
What would the same usage cost through an API?
For our local Qwen3.8-Flash-Next variants, we estimate approximately €457 in equivalent API charges using OpenRouter’s published Qwen3.8 Flash rates without a cache discount. This is a comparison, not a bill we paid or a proven saving. The online model, local quantized variants, results, pricing, and cache behavior may differ. Other text models, images, videos, hardware, electricity, and development time are excluded from this figure.
Our image and video experiments could add to that comparison, but our video inventory still needs further duplicate checks, so we are not presenting a definitive combined total.
What comes next for ObviousIdea?
We are continuing to build and test local AI workflows, from autonomous agents to image and video tools. Expect to hear more as we learn what works and how these experiments may shape future ObviousIdea products and interfaces. Features and availability will be announced when they are ready.
Share your experience with local AI
Are you an AI enthusiast, developer, creator, or privacy-conscious user exploring local AI? We would love to hear what you are building, the challenges you have encountered, and the features you would find useful. Contato ObviousIdea and share your experience.
Source: ObviousIdea local usage records, November 2–15, 2026. API comparison uses input and output list prices without a cache discount and an illustrative exchange rate of €1 = $1.1165.
