将 PDF 和图片转换为 10 种文档或数据格式,并提供 OCR 与版面分析控制。
设计与多媒体
pdf-online
试用Parse PDF, image, Word, or PPT files with SoMark and publish the result to Feishu, DingTalk, or Notion as editable documents, spreadsheets, record tables, or databases. Also use when an exact matching SoMark Markdown-and-JSON pair must be imported into one or more of these platforms.
它能做什么
Parse PDF, image, Word, or PPT files with SoMark and publish the result to Feishu, DingTalk, or Notion as editable documents, spreadsheets, record tables, or databases. Also use when an exact matching SoMark Markdown-and-JSON pair must be imported into one or more of these platforms.
技能文档
PDF Online
Route a publishing request to the existing platform adapter. This file is only the entry point: do not reproduce, merge, or improvise platform conversion rules here.
Select the platform
- Feishu, Lark, a Feishu/Lark destination URL, cloud document, Feishu Sheet, or Bitable request: read platforms/feishu/SKILL.md completely and follow it.
- DingTalk, Alidocs, a DingTalk destination URL, DingTalk document, online spreadsheet, or AI Table request: read platforms/dingtalk/SKILL.md completely and follow it.
- Notion, a Notion destination URL, Notion page, or Notion database request: read platforms/notion/SKILL.md completely and follow it.
- When exactly one platform is implied, route directly. When the platform is genuinely ambiguous, ask only which of Feishu, DingTalk, or Notion to use.
- When the user explicitly requests multiple platforms, load only those platform modules and execute each independently.
Treat the selected platform directory as that adapter's skill directory.
Resolve every scripts/..., references/..., and `` path in its
instructions relative to that platform directory, never relative to this root.
Read platform-specific references only when the selected module directs it.
Do not load or mix rules from an unselected platform.
Shared parsing boundary
- Never invoke SoMark more than once for the same source in one task.
- For every raw PDF, image, Word, or PPT source, the Agent must invoke the
separately installed official
somark-document-parserSkill and obtain its Markdown-and-JSON result before running a platform adapter. The adapters do not contain the parser and must not search parent or sibling directories for a parser script. - For one platform, give that exact pair to the selected adapter. For multiple platforms, invoke the parser Skill once and give the same pair to every selected adapter. A retry in one platform must reuse the same pair.
- When the user explicitly supplies an exact matching Markdown-and-JSON pair, do not invoke SoMark and do not search for replacements or historical results.
Execution boundary
The selected platform module is authoritative for target-type routing, conversion, publishing, verification, recovery, degradation, and user-visible status. Keep its scripts and safety rules unchanged. A success or failure on one requested platform does not authorize creating a replacement target or rerunning SoMark for another platform.
相关技能
通过 xparse-cli 把 PDF、图片、Office、HTML、OFD 等文档转为结构化的 Markdown 或 JSON。
Convert PDF documents to PowerPoint presentations via GoAI API. Use when the user asks to convert PDF to PPT, turn a PDF into slides, make a presentation fro...
Design or implement a paper-management workflow built on a Feishu bot plus OpenClaw. Use when the user wants to ingest papers from Feishu messages, save PDFs...
Convert unstructured documents into LLM-ready structured data. Supports PDF, Word, PPT, and images; extracts paragraphs, formulas, tables, charts, and other...
Convert Excel (.xlsx) sign-in sheets and rosters to print-ready A4 PDFs with Chinese font support. Use when the user needs to (1) convert an Excel table/rost...