She Walks: LTX 2.5 First/Last-Frame Techniques for a Moving AI Presenter

How we made our AI presenter walk between sets in one continuous shot on a single RTX 5090: the walk-set plate design, guide strength 0.7 to 0.5, the 47% midpoint problem, generate-long-cut-short, chroma key at 0.28, and the honest numbers including the advice of ours that testing killed.

August 30, 2026 · 7 min · Aillex / DIY AI
Eros 10Eros local text to audio-video

Eros (10Eros): One Local Model That Makes Video AND Audio

Eros is a baked LTX-2.3 variant that generates synced audio and video from a prompt in one model. It’s wild, but it writes its own soundtrack, so it can’t dub your voice. Here’s the honest picture.

July 24, 2026 · 2 min · Aillex / DIY AI
LTX-2.3 local AI video

LTX-2.3: The Local AI Video Model I Actually Use

LTX-2.3 is the fast, local, image-to-video model that became our channel’s workhorse, 30fps, native portrait, faster-than-realtime. Here’s what it’s great at, where it slips, and why we moved talking to a specialist.

July 24, 2026 · 2 min · Aillex / DIY AI
Wan 2.2 local benchmark

Wan 2.2 Locally: I Benched 5B vs 14B on One Card

Everyone’s arguing about Wan 2.2. So we ran it locally, honestly, the fast-and-fits 5B, the heavy two-expert 14B, the VRAM wall, and the decode that wedges. Here are the real numbers.

July 24, 2026 · 2 min · Aillex / DIY AI
As an Amazon Associate, this site earns from qualifying purchases. Referral links are always disclosed.