Breaking

DepthFlow, an Open-Source Tool Turning Still Images Into 3D Parallax Videos

July 5, 2026 at 09:55 EDT

Media Generation
Open Source

DepthFlow, an open-source tool that converts static images into videos with a 3D parallax effect, is drawing attention. Developed by BrokenSource, it is released on GitHub under the AGPL-3.0 license, with commercial use also permitted.

Open Source · DepthFlow

One Still Image → 3D Parallax Video, Free and Local

DepthFlow applies a depth-map parallax effect to turn a single photo into camera-motion animation — running locally, with no watermarks and commercial use permitted. It's positioned as a free alternative to commercial services like ImmersityAI.

Max render resolution at 50fps on an RTX 3060

<4GB

VRAM used in a low-spec example workflow

1.4k

GitHub stars (as of July 2026)

Frame throughput reported in the field

Fast processing even at high frame counts — the taller column is the same clip at 4× the frames.

baseline frames

240

frames @ 120fps

How it works

Still image

A single photo input

→

Depth map

SOTA models: Depth Anything V2/V3, Depth Pro

→

Parallax render

Ray-marching GLSL shader + TurboPipe → FFmpeg

→

2.5D / 3D clip

Zoom · Dolly · Circle · H/V presets, seamless loop

Why developers like it

Free, local, no watermarks, commercial use OK (AGPL-3.0)
Runs on low VRAM and stays fast
Strong fit with Stable Diffusion / ComfyUI and LTX Video
Python, Web UI, ComfyUI node, and Android app

The caveats

Effect quality depends on depth-map accuracy
Inaccurate depth weakens the parallax
GPU performance matters significantly
Setup best done via ComfyUI Manager

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.