帧级 OCR 识别
对单帧图像或视频关键帧执行 OCR，返回可结构化文本，适合字幕、单据与画面文字提取。
- Capability ID: video.engine.vision.ocr_frame
- Playground trial available
- Publisher: ONLINESVR
# 帧级 OCR 识别
从图片或视频帧中识别文字，输出结构化结果。
## 适用场景
- 视频字幕/画面文字提取
- 单据与标签 OCR
- 视觉流水线前置步骤
## 试用
可在 Playground 发起试用（受试用额度与白名单约束）。Frequently Asked Questions
How do I run this Capability?
Use Try for free (Playground) when enabled, or submit via Gateway after sign-in.
Is the Store listing the same as execution permission?
No. Published listing is discovery only; each run still checks identity, entitlement, schema, and release.