Generate video dari teks (text-to-video) pakai model Wan 2.2 — output MP4 tanpa watermark/login. Catatan: rate-limited (kuota GPU), URL video ephemeral.
promptstringRequiredText prompt describing the video to generate
imagestringOptionalOptional source image URL to animate (image-to-video). When provided, the model animates this image guided by `prompt`; output dimensions auto-fit the image. Omit for pure text-to-video
modelenumOptionalModel. Only wan (Wan 2.2 5B, fast 4-step) is available. Default wan
wandurationnumberOptionalVideo length in seconds (1-8). Default 2
widthnumberOptionalOutput width px, multiple of 32 (256-1024). Default 704
heightnumberOptionalOutput height px, multiple of 32 (256-1024). Default 704
seednumberOptionalSeed for reproducibility. Default random
negativePromptstringOptionalThings to avoid in the video. Has a sensible default
stepsnumberOptionalInference steps (1-8). More steps = higher quality but slower. Default 4
guidancenumberOptionalGuidance scale (0-5). Higher = follow prompt more strictly. Default 0 (recommended for the fast model)
Chatex chat — OpenAI-compatible chat completions.
AI search assistant dengan web real-time — OpenAI-compatible, support streaming, return sumber referensi.
Microsoft Copilot — chat completion OpenAI-compatible, support streaming SSE
Dolphin 24B uncensored chat — chat completion OpenAI-compatible, support streaming SSE
LLM-ready markdown extraction with anti-bot escalation. 3-tier: fast (direct), magic (browser stealth), worker-chain (CF Worker IP rotation).
Anonymous AI image generation — text-to-image, no auth.