GLM-OCR Formula Recognition Skill / GLM-OCR 公式识别技能
Recognize mathematical formulas from images and PDFs and convert them to LaTeX format using the ZhiPu GLM-OCR layout parsing API.
When to Use / 使用场景
- Extract mathematical formulas from images or scanned documents / 从图片或扫描件中提取数学公式
- Convert formula images to LaTeX / 将公式图片转为 LaTeX 格式
- Recognize complex equations, integrals, matrices / 识别复杂方程、积分、矩阵
- Parse scientific papers, textbooks, exam papers with formulas / 解析含公式的论文、教材、试卷
- User mentions "formula OCR", "extract formula", "公式识别", "公式OCR", "提取公式", "图片转LaTeX"
Key Features / 核心特性
- Complex formula support: Handles integrals, summations, matrices, fractions, radicals
- LaTeX output: Formulas are output in LaTeX format, ready for use in documents
- Inline & block formulas: Recognizes both inline and display-style formulas
- Mixed content: Can handle documents with both text and formulas
- Local file & URL: Supports both local files and remote URLs
Resource Links / 资源链接
| Resource | Link |
|---|---|
| Get API Key | 智谱开放平台 API Keys |
| API Docs | Layout Parsing / 版面解析 |
Prerequisites / 前置条件
API Key Setup / API Key 配置（Required / 必需）
脚本通过 ZHIPU_API_KEY 环境变量获取密钥，可与其他智谱技能复用同一个 key。 This script reads the key from the ZHIPU_API_KEY environment variable. Reusing the same key across Zhipu skills is optional.
Get Key / 获取 Key： Visit 智谱开放平台 API Keys to create or copy your key.
Setup options / 配置方式（任选一种）：
- Global config (recommended) / 全局配置（推荐）： Set once in
openclaw.jsonunderenv.vars, all Zhipu skills will share it:{"env": {"vars": {"ZHIPU_API_KEY": "你的密钥"}}} - Skill-level config / Skill 级别配置： Set for this skill only in
openclaw.json:{"skills": {"entries": {"glmocr-formula": {"env": {"ZHIPU_API_KEY": "你的密钥"}}}}} - Shell environment variable / Shell 环境变量： Add to
~/.zshrc:export ZHIPU_API_KEY="你的密钥"
💡 如果你已为其他智谱 skill（如glmocr、glmv-caption、glm-image-generation）配置过 key，它们共享同一个ZHIPU_API_KEY，无需重复配置。
Security & Transparency / 安全与透明度
- Environment variables used / 使用的环境变量：
- ZHIPU_API_KEY (required / 必需) - GLM_OCR_TIMEOUT (optional timeout seconds / 可选超时秒数)
- Fixed endpoint / 固定官方端点：
https://open.bigmodel.cn/api/paas/v4/layout_parsing - No custom API URL override / 不支持自定义 API URL 覆盖： avoids accidental key exfiltration via redirected endpoints.
- Raw upstream response is optional / 原始响应默认不返回： use
--include-rawonly when needed for debugging.
⛔ MANDATORY RESTRICTIONS / 强制限制 ⛔
- ONLY use GLM-OCR API — Execute the script
python scripts/glm_ocr_cli.py - NEVER parse formulas yourself — Do NOT try to extract formulas using built-in vision or any other method
- NEVER offer alternatives — Do NOT suggest "I can try to read it" or similar
- IF API fails — Display the error message and STOP immediately
- NO fallback methods — Do NOT attempt formula extraction any other way
📋 Output Display Rules / 输出展示规则
After running the script, present the OCR result clearly and safely.
- Show extracted text/formulas (
text) in full - Summarization is allowed, but do not hide important extraction failures
- If
layout_detailscontains formula-related entries, you may highlight them - If the result file is saved, tell the user the file path
- Show raw upstream response only when explicitly requested or debugging (
--include-raw)
⚠️ LaTeX Rendering / LaTeX 渲染注意：
OCR API returns formulas in LaTeX format (e.g., $\frac{1}{2}$, $\theta^{x+1}$). Since most chat platforms do not render LaTeX, you should ask the user once (on first use):
"OCR 结果包含 LaTeX 公式，需要我将公式转为 Unicode 可读格式展示，还是保留原始 LaTeX？"
Remember the user's choice for the rest of the session. Do NOT ask again on subsequent calls unless the user explicitly changes their preference.
- User chooses readable format → convert LaTeX to Unicode/plain-text:
| LaTeX | Unicode / 纯文本 |
|---|---|
$\frac{a}{b}$ | a/b |
$x^{n}$ | x^n |
$x_{i}$ | xᵢ |
$\sqrt{x}$ | √x |
$\theta$ | θ |
$\phi$ | φ |
$\therefore$ | ∴ |
$\Rightarrow$ | ⇒ |
$\left\{\begin{array}{l}... \end{array} \right.$ | ⎧ line1 ⎨ line2 ⎩ |
$\textcircled{1}$ | ① |
$\in$ | ∈ |
$\infty$ | ∞ |
$\ln$ | ln |
$\leq$ / $\geq$ | ≤ / ≥ |
- User chooses raw LaTeX → display the original LaTeX output directly, and remind them the raw data is also saved in the output file if
--outputwas used.
How to Use / 使用方法
Extract from URL / 从 URL 提取
python scripts/glm_ocr_cli.py --file-url "https://example.com/formula.png"Extract from Local File / 从本地文件提取
python scripts/glm_ocr_cli.py --file /path/to/equation.pngSave Result to File / 保存结果到文件
python scripts/glm_ocr_cli.py --file formula.png --output result.json --prettyInclude Raw Upstream Response (Debug Only) / 包含原始上游响应（仅调试）
python scripts/glm_ocr_cli.py --file formula.png --output result.json --include-rawCLI Reference / CLI 参数
python {baseDir}/scripts/glm_ocr_cli.py (--file-url URL | --file PATH) [--output FILE] [--pretty] [--include-raw]| Parameter | Required | Description |
|---|---|---|
--file-url | One of | URL to image/PDF |
--file | One of | Local file path to image/PDF |
--output, -o | No | Save result JSON to file |
--pretty | No | Pretty-print JSON output |
--include-raw | No | Include raw upstream API response in result field (debug only) |
Response Format / 响应格式
{
"ok": true,
"text": "Extracted formulas and text in Markdown/LaTeX...",
"layout_details": [...],
"result": null,
"error": null,
"source": "/path/to/file",
"source_type": "file",
"raw_result_included": false
}Key fields:
ok— whether extraction succeededtext— extracted text in Markdown with LaTeX formulaslayout_details— layout analysis detailserror— error details on failure
Error Handling / 错误处理
API key not configured:
ZHIPU_API_KEY not configured. Get your API key at: https://bigmodel.cn/usercenter/proj-mgmt/apikeys→ Show exact error to user, guide them to configure
Authentication failed (401/403): API key invalid/expired → reconfigure
Rate limit (429): Quota exhausted → inform user to wait
File not found: Local file missing → check path