Borrowing it
Nothing to install: this file belongs to tinygone/unlimited-ocr-mcp-server. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/tinygone/unlimited-ocr-mcp-server/master/.claude/skills/uocr/SKILL.mdgit clone --depth 1 https://github.com/tinygone/unlimited-ocr-mcp-serverWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tinygone/unlimited-ocr-mcp-server/uocr)<a href="https://agentmods.dev/skills/tinygone/unlimited-ocr-mcp-server/uocr"><img src="https://agentmods.dev/badge/skills/tinygone/unlimited-ocr-mcp-server/uocr/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tinygone/unlimited-ocr-mcp-server/uocr"><img src="https://agentmods.dev/badge/skills/tinygone/unlimited-ocr-mcp-server/uocr.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00084 | $0.01278 |
| Opus 5.5 | $0.00034 | $0.00511 |
| Sonnet 5.5 | $0.00017 | $0.00256 |
| Haiku 4.5 | $0.00008 | $0.00128 |
Grade A, and why
uocr scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- 验证:`curl http://127.0.0.1:10000/v1/models` 应返回模型列表 How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
UOCR Skill
⚠️ 前置条件(必须满足才能用)
本 Skill 调用的 mcp__uocr__ocr_image / mcp__uocr__ocr_multi 工具,背后依赖一个本地 OCR 服务,必须先启动:
- OCR 服务必须在线:另一个终端运行
python serve_ocr.py,监听http://127.0.0.1:10000- 验证:
curl http://127.0.0.1:10000/v1/models应返回模型列表 - 服务未启动时调用会卡住直到超时(默认 1800s)
- 验证:
- 环境变量(首次下载模型需要):
HF_ENDPOINT=https://hf-mirror.com(国内) - conda 环境:需激活
uocr环境(含 PyTorch + transformers)
详见项目根目录 README.md。
技能目标
使用 uocr MCP server 识别图片中的文字内容,并在拿到识别结果后,按用户原始诉求继续处理(总结、翻译、结构化、回答问题等)。
核心原则:不要在调用 OCR 工具前猜测图片内容。
触发条件
当任务中出现以下任一情况时,必须使用本技能:
- 用户提供了图片(截图、扫描件、拍照件、文档图片等)
- 用户要求"识别这张图""看看图里写了什么""OCR 一下""读取图片"等
- 任务需要从图片中提取文字信息(例如发票、合同、表单、UI 截图等)
可用工具(由 uocr MCP server 提供)
1. ocr_image — 识别单张图片
参数:
image_path(推荐):本地图片绝对路径,例如D:/docs/page.pngimage_base64:图片 base64 编码(不带data:前缀);没有路径时使用prompt:识别任务描述,默认"document parsing."
调用示例:
{
"image_path": "D:/aiworkspace/uocr-workspace/inputs/1.jpg"
}
2. ocr_multi — 识别多张图片(多页文档)
参数:
image_paths:图片路径列表,例如["page1.png", "page2.png"]prompt:任务描述,默认"Multi page parsing."
调用示例:
{
"image_paths": [
"D:/docs/page1.png",
"D:/docs/page2.png"
]
}
多张图会一起识别,能更好地保留跨页上下文,适合多页扫描件、PDF 转出的多张图。
执行流程
-
判断工具:
- 单张图 →
ocr_image - 多张图(≥2)→
ocr_multi
- 单张图 →
-
构造参数:
- 优先使用
image_path(本地绝对路径) - 路径需为绝对路径,正斜杠
/或双反斜杠\\均可 - 只有在拿不到本地路径时,才退而使用
image_base64(不带data:前缀)
- 优先使用
-
调用工具:调用对应的
mcp__uocr__ocr_image或mcp__uocr__ocr_multi。 -
处理结果:
- 工具返回的是纯文本识别结果(可能附带一段结构化 JSON)
- 不要把原始 JSON 直接丢给用户,应整理成易读的形式
- 拿到结果后,继续按用户的原始诉求处理:
- 仅识别 → 给出识别文字
- 要求总结 → 总结
- 要求翻译 → 翻译
- 要求结构化 → 输出表格 / JSON / Markdown
- 要求回答问题 → 基于识别内容作答
-
格式建议:
- 按图片版面结构组织(标题、正文、表格、底部按钮等)
- 繁体内容可附简体摘要
- 主动给出后续可选操作(翻译、结构化导出、对比其他截图等)
注意事项
- ❌ 不要在调用 OCR 工具前猜测或描述图片内容
- ❌ 不要把工具返回的原始 JSON 直接作为最终答复
- ✅ 单图用
ocr_image,多图用ocr_multi - ✅ 优先
image_path,必要时才用image_base64 - ✅
image_base64必须去掉data:image/...;base64,前缀 - ✅ 拿到结果后必须做二次加工,不能止步于返回原文
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 115 lines · 84 tokens per session scan A 8e68bb910eaf
uocr is a skill published in the GitHub repository tinygone/unlimited-ocr-mcp-server (0 stars, last pushed 3mo ago), licensed MIT. It adds 84 tokens to every session and 1,278 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
uocr compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| uocrtinygone/unlimited-ocr-mcp-server | Skill | 0 | 3mo ago | MIT | A | 84 | Claude Code |
| orbit-notionnexu-io/open-design | Skill | 107k | today | Apache-2.0 | A | 117 | Claude Code, Codex |
| instrument-data-to-allotropeanthropics/knowledge-work-plugins | Skill | 35k | yesterday | Apache-2.0 | A | 123 | Claude Code |
| apify-youtube-transcript-scrapersickn33/agentic-awesome-skills | Skill | 50k | today | MIT | A | 48 | Claude Code |
| report-to-pdfasgeirtj/system_prompts_leaks | Skill | 71k | yesterday | CC0-1.0 | A | 17 | Claude Code, Codex |
| legal-holdanthropics/claude-for-legal | Skill | 10.0k | 3d ago | Apache-2.0 | A | 68 | Claude Code |
| artifact-template-minimal-letterheadasgeirtj/system_prompts_leaks | Skill | 71k | yesterday | CC0-1.0 | A | 54 | Claude Code, Codex |