Kinoloom: open-source ComfyUI film studio renders seamless LTX 2.5 takes on 8GB card
Dry-Acanthisitta-652 · reddit · 2026-10-30
A developer released Kinoloom, an MIT-licensed film studio built on ComfyUI that renders everything locally — demonstrated on an RTX 3070 (8GB) with 64GB RAM.
Features
- Video studio: given a start image and prompt per piece, LTX 2.5 renders them as one continuous take with sound
- Image studio: stills and start pictures
- Movie mode: cast characters and locations, write the story with any LLM, then render and review
How seamless takes work: the end of each piece is passed to the next as latents before pixel conversion, with the whole take decoded in one pass.
Timings on RTX 3070: LTX 2.5 ≈6 min per 5s 1080p clip warm; MiniMax H3 20-22 min (weights not downloaded due to license restrictions in EU/UK/US/South Korea); Qwen image 2.1 fast 38-78s per still.
Requirements: Windows 11, 8GB+ NVIDIA GPU (GTX 16/RTX 20 or newer), 64GB RAM (32GB reportedly works for LTX 2.5). INSTALL.bat sets everything up; seamless chains remain experimental.
Repo: github.com/harlixay7/Kinoloom — an entry for the Comfy Open Source Higgsfield challenge.
More from coding & agent
- GitHub Copilot with Opus 5.5 beats Claude Code with the same model, dev claims — DimitrisPapail · 2026-10-30
- Two agent runs crashed on their own plan files—7 self-repair rules to save the whole run — petrusenko_max · 2026-10-30
- NVIDIA releases open Nemotron 3.5 Embed, Rerank and Parse models to cut agent token costs — NVIDIAAI · 2026-10-30
- The Infinite Work Machine: buyer bots vs seller bots turn AI efficiency gains into more work — Ididit-forthecookie · 2026-10-30
- Seroter's reading list: agent identity management, AI code review fails, skills as products — rseroter · 2026-10-30
- Supabase shares Multigres build story and opens early access signups — dshukertjr · 2026-10-30