Yes — the technique you’re describing is generally called multi-frame super-resolution, often shortened to super-res. It works by aligning multiple frames and combining their slightly different detail and noise patterns to reconstruct a better still than a single frame alone.

In practice, tools exist, but they’re less common as simple consumer apps because results depend heavily on the footage and on careful tuning. Python-based workflows are available, and some camera systems now include similar ideas internally. Astronomical imaging software can also be useful, since stacking and alignment are common there; techniques such as drizzle may help recover a bit more detail when there are small frame shifts.

The main limitation is that this is not magic: it works best with stable scenes, good source quality, small natural sub-pixel shifts, and enough frames. If the video is heavily compressed, blurry, or the subject is moving, gains may be limited.

So: yes, this is a real method, but it’s usually called multi-frame super-resolution / frame stacking rather than ordinary upscaling.