Discofork
Repository brief

google-ai-edge/mediapipe

Read the upstream summary on the left, browse the cached forks below it, and load each fork comparison into the right-hand panel.

Cached analysis
cached 2026-03-30T20:04:17.607Z
1mo ago

google-ai-edge/mediapipe

MediaPipe is a large, active open source repo for cross-platform machine learning on live and streaming media. It supports Android, iOS, web, desktop, edge devices, and IoT, and includes both higher-level MediaPipe Solutions and the lower-level MediaPipe Framework.

GitHub
Loading tags...
Stars34,416
Forks5,869
Default branchmaster
Last pushed2026-03-28T04:55:34Z
Recommended shortcuts

Jump straight into Discofork's strongest cached fork picks, or open a compare view in one click.

Forks

Choose a fork to inspect

10 of 10 fork briefs
Selected

Prefer upstream unless you specifically need this older hand-gesture-focused snapshot; the fork looks specialized, heavily outdated, and costly to keep compatible.

Prefer this fork only if Windows/VS2019 support is the primary constraint. If you need current MediaPipe features, active maintenance, or broad cross-platform support, upstream is the better default.

Choose this fork only if you specifically want a face-mesh demo/prototype and accept major staleness. For most adopters, upstream MediaPipe is the safer choice because it is far more current, broader in scope, and actively maintained.

Choose this fork only if you specifically need its CUDA-oriented customization and can tolerate a stale, heavily diverged codebase. For most adopters, upstream is the safer default because it is much newer and retains broader platform and solution coverage.

Choose this fork only if you want a narrow Linux-specific MediaPipe base and are prepared to own the maintenance. If you need current MediaPipe capabilities, cross-platform support, or the full Tasks/solutions stack, upstream is the better choice.

Prefer this fork if you want MediaPipe specialized around genai inference and API refactoring, and you can tolerate losing some upstream breadth and carrying your own integration risk. Prefer upstream if you need the full, actively maintained MediaPipe surface area and lower maintenance overhead.

Prefer this fork only if you need an old, preserved MediaPipe snapshot or its local workflow/doc changes. If you want active maintenance, current APIs, or broad example/test coverage, upstream is the better choice.

Choose this fork only if its older face/geometry experiments are the point. For general MediaPipe adoption, upstream is the better default because this fork is stale, heavily diverged, and has removed a lot of supporting assets.

Adopt this fork only if OSC hand-tracking is the core requirement. If you need current MediaPipe Tasks, broader platform support, or an actively maintained codebase, upstream is the better base.

Choose this fork if your priority is OpenVINO/OVMS integration and related production workflows. Choose upstream if you need the widest MediaPipe platform coverage, especially newer web or general-purpose features.