SadTalker
Free.ai (self-hosted)
·
video
·
~30000 tokens per video clip
·
3.2
from 79 users of this category
SadTalker is a video generation model built by OpenTalker. Strongest at Avatar explainers, voice-over to video, talking-head ads, dubbed content.. Self-hosted on Free.ai GPUs — runs free against your daily token pool (30,000 tokens per ~5s clip). Released under Non-commercial (BFM) — commercial use permitted on Free.ai.
Use via API
OpenAI-compatible REST API. Generate a key and call this model in seconds.
curl -X POST https://api.free.ai/v1/video/generate/ \
-H "Authorization: Bearer sk-free-..." \
-H "Content-Type: application/json" \
-d '{"model":"sadtalker","prompt":"your prompt here"}'Similar models
Frequently Asked Questions
SadTalker is one of the AI video models available on Free.ai. Use the form on this page to generate a clip; below the fold you will find sample prompts, pricing, and alternatives.
On Free.ai, video generation tops out at 6 seconds per clip across all video models. To extend, use the Extend button on the full /video/ tool — it captures the last frame and continues from there.
16:9 landscape (YouTube, desktop), 9:16 portrait (TikTok, Reels, Shorts), and 1:1 square (Instagram feed). Pick the ratio in the form before you generate so the model frames the shot for that shape.
720p HD by default. Some premium models support up to 1080p — check the Compare card. You can upscale any output in /video/upscaler/ after the fact.
Strong on natural motion (people walking, cars driving, water, weather). Weaker on intricate hand interaction and fine text. Use camera-direction hints (slow pan, dolly forward) for cleaner motion.
Free.ai grants commercial use of generated videos to you regardless of model. The model itself may be Apache 2.0, MIT, or a vendor-specific license — see the model reference link below.
SadTalker runs on our own GPUs at 30,000 tokens per clip, flat: a render costs what it costs because of model load and diffusion steps, not clip length, and 2,500 per second only takes over past 12 seconds. That is the entire 30,000-token daily free pool, so a free account gets one clip every two days. Paid token packs start at $5 → 200,000 tokens and remove the cap.
Yes — use /batch/ for up to 100 prompts. Each clip is queued and the result lands in /account/?tab=history when ready.
Yes — POST your prompt to /v1/video/generate/ with model="SadTalker" (or use the slug on this page). See /api/ for the full endpoint reference plus auth, error codes, and SDK snippets.
We do not store your prompts beyond the share window (24h anon, 7d signed-in). We do not train on your inputs. Self-hosted models run on Free.ai-owned GPUs.
Yes — failed jobs are auto-refunded to the original payment source (daily pool or paid tokens) the same day. Email contact@free.ai if a refund did not appear.
Typical render 30–120 seconds for a 4-second clip; premium models often finish in 15–45 seconds. Use the close-tab-safe queue button on /video/ — your result lands in /account/?tab=history when done, with optional email.