Models

SDXL-based Cyber-XL on local RTX pool. ~25s/image.

FLUX 2 Klein 4B on local RTX pool. ~18s/image.

High-fidelity text-to-image. Photorealistic and stylized. Up to 1536px.

Prompt-guided image editing on 1, 2, or 3 input images. RunPod serverless on 24GB Pro.

Fast turbo text-to-image. Good for thumbnails, drafts, and high-volume use.

Fast Z-Image Turbo on local RTX pool. ~12s/image.

Fast text LLM with strong tool/function-calling support. 32K context, ideal for agents.

Multimodal LLM (Qwen3-VL-30B-A3B-Instruct, MoE 3B active). Best for text reasoning, OCR, structured extraction. 32K context. NOTE: deploymen…

Sanitized vision LLM tuned for public-facing applications. Lower latency, same vision capability as Jessi 72B, content filters enabled.

Long-form narrative LLM. Plot reasoning, character consistency, scene blocking. Text-only.

Image-to-video and speech-to-video. 5-10 second clips at 720p, 24fps.