.ocr_imageCall via POST http://127.0.0.1:3000/ocr_image (JSON params as the body) or WebSocket.
| Param | Type | Required | Default | Description |
|---|---|---|---|---|
image |
string | ✓ |
{ texts, language }:识别文本数组(含置信度与坐标)与识别语言。
Every response is wrapped in the standard envelope: { "status": "ok", "retcode": 0, "data": ... }. The table below describes data.
| Field | Type | Required | Description |
|---|---|---|---|
texts |
object[] | ✓ | 识别出的文本块 |
language |
string | ✓ | 识别语言 |
MCP-capable clients can discover and call this action directly — no hand-written HTTP. See MCP.