Grok 文本、图片、语音与视频

Grok 文本通常使用 OpenAI 兼容协议；图片、视频和语音使用当前网关提供的媒体路由。分组必须开启对应媒体能力。

文本

curl "$BASE_URL/v1/chat/completions" -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" -d '{"model":"GROK_MODEL","messages":[{"role":"user","content":"你好"}]}'

图片

curl "$BASE_URL/v1/images/generations" -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" -d '{"model":"grok-imagine","prompt":"一只戴眼镜的橘猫，电影海报风格"}'

图片编辑使用 /v1/images/edits，请求体和文件上传方式见 OpenAI 生图。

语音

TTS

curl "$BASE_URL/v1/tts" -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" -d '{"model":"grok-voice-latest","input":"欢迎使用瑞科 API","voice":"default"}' --output speech.mp3

STT

curl "$BASE_URL/v1/stt" -H "Authorization: Bearer $API_KEY" -F "file=@speech.mp3" -F "model=grok-voice-latest"

Realtime

GET /v1/realtime?model=grok-voice-latest

这是 WebSocket，不是普通 HTTP POST。客户端需要完成 WebSocket 升级并按 xAI Realtime 事件格式发送音频。

视频

curl "$BASE_URL/v1/videos" -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" -d '{"model":"grok-imagine-video","prompt":"海边日落，镜头缓慢推进","duration":4,"resolution":"720p"}'

视频流程：POST /v1/videos 创建任务，GET /v1/videos/{id} 查询状态，完成后 GET /v1/videos/{id}/content 下载。编辑和续写分别使用 /v1/videos/edits、/v1/videos/extensions。