Reddit user seeks an autonomous agent that builds and tunes ComfyUI video workflows end-to-end
Hopeful-Election-783 · reddit · 2026-09-19
A Reddit user is asking whether an autonomous agent for local AI workflows already exists: feed it a script, have it split into scenes, detect which local image/video models are available and pick the right ones, and even research speedups, modify workflows, test, and iterate on its own.
The vision is a high-level task like "turn this script into a video" handled entirely by the agent — model selection, workflow construction, generation, troubleshooting, node tuning, and refinement — using ComfyUI, local LLMs, or other agent frameworks. No mature solution has emerged yet.
More from coding & agent
- AI agent books a United flight change on its own as 'do it for me' bar rises — armand_ruiz · 2026-09-19
- Clinical agent run costs $0.0047: 99 calls, 143K tokens with Jev in a triage workflow — MaziyarPanahi · 2026-09-19
- Alchemy IaC adds secret manager support for Doppler, Infisical and more — samgoodwin89 · 2026-09-19
- Swarms to ship MCP Deployer in v16, turning any agent into an MCP server in 5 lines — KyeGomezB · 2026-09-19
- Letting agents deploy your app then browse it like real users: super high-level fuzzing — lucasmeijer · 2026-09-19
- Prompting rule: ban fallbacks so AI models can't cheat around core work — Daniel_Farinax · 2026-09-19