dsh-tool-ocr
Standalone ocr tool plugin for DeepSeek Harness (dsh) — local image text recognition via the newbee-ocr (nbocr) engine, for models without vision input
AI Analysis
核心用途是为纯文本模型提供本地离线图片文字识别能力。适合无法使用多模态模型但又需要处理图片中文字提取任务的用户。必要条件是需编译并配置本地 nbocr 二进制文件。
Install
This plugin has no verified bundle, or compatibility checks failed. Read the repository notes first. Read the full README ↗
README
Read the full README ↗Configuration
| Field | Default | Meaning |
|---|---|---|
command | — (required) | The nbocr executable: absolute path or PATH-resolved name. |
args | [] | Extra arguments before the nbocr subcommand (no shell). |
env | {} | Extra environment entries. |
language | chinese | Recognition model/language alias. |
detModel | v6-tiny | Detection model tier. |
modelsDir | '' | Model directory for non-embedded models; empty uses embedded models. |
maxImageBytes | 26214400 | Largest accepted image in bytes. |
maxOutputBytes | 2000000 | Largest collected engine stdout in bytes. |
maxTextChars | 12000 | Largest recognized text returned in text. |
timeoutMs | 600000 | Tool-call timeout budget in ms. |
Tool usage
ocr { path | attachment_id, action?, include_boxes?, table?, max_text_chars? }
The model should call ocr with path for image files, or with attachment_id when the conversation already contains an image (shown as [image attachment ]). status probes readiness without engine work; check runs a real 1x1 probe recognition end to end.