Oak & Fir Ventures
Applied AI for motion and sound.
Oak & Fir Ventures is an independent software company building products and applications that understand and act on continuous motion and sound, from capture to response.
Why motion and sound
- Language is a tiny sliver of how humans interact with the world. Every movement, every sound, every shift in rhythm carries meaning, and almost none of it ever becomes words. Yet language still sits at the center of most AI systems, and of how most agents plan and act.
- Motion and sound don't follow language's rules. They are continuous, noisy, and arrive together: a movement makes a sound, and a sound sets something in motion. The moments that matter can last mere milliseconds, and any response that arrives late is the wrong one.
- Perception is getting easier. Timing is not. Multimodal models can now describe some of what they see and hear, usually after the fact and as a summary. We build systems and agents that keep pace with the moment, so they can change what happens next.
What we're building
Motion
Systems that track how things move in ordinary video, frame by frame, and turn it into precise, repeatable results.
Sound
Systems that follow the timing, structure, and texture of sound as it happens, even in noisy, real-world environments.
Data
Scalable pipelines that capture real-world motion and sound, verify exactly what happened, and make it ready for training.
How we work
- Capture: ordinary hardware. Everyday cameras and microphones, including a phone. No add-on sensors or specialized equipment, so it scales without new hardware.
- Inference: device and cloud. Time-critical work runs on the device and heavier analysis on GPUs in the cloud, both built to run continuously at low cost.
- Training: verified ground truth. Every result is measured against what actually happened, so models learn from fact, not guesswork.
Contact
For partnerships or inquiries: hello [at] oakandfirventures.com