Quick behavior:<\/p>\n The tool is only registered when OpenClaw can resolve a PDF-capable model config for the agent:<\/p>\n If no usable model can be resolved, the Input notes:<\/p>\n Reference notes:<\/p>\n Native mode is used for provider Native mode limits:<\/p>\n Fallback mode is used for non-native providers.<\/p>\n Flow:<\/p>\n Fallback details:<\/p>\n “`json5 theme={“theme”:{“light”:”min-light”,”dark”:”min-dark”}} Multiple PDFs:<\/p>\n “`json theme={“theme”:{“light”:”min-light”,”dark”:”min-dark”}} PDF Tool PDF tool pdf analyzes one or more PDF document […]<\/p>\n","protected":false},"author":0,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[],"class_list":["post-644","post","type-post","status-publish","format-standard","hentry","category-docs"],"_links":{"self":[{"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/posts\/644","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/comments?post=644"}],"version-history":[{"count":2,"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/posts\/644\/revisions"}],"predecessor-version":[{"id":699,"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/posts\/644\/revisions\/699"}],"wp:attachment":[{"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/media?parent=644"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/categories?post=644"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/pa.yingzhi8.cn\/index.php\/wp-json\/wp\/v2\/tags?post=644"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}pdf<\/code> analyzes one or more PDF documents and returns text.<\/p>\n\n
pdf<\/code>) or multi (pdfs<\/code>) input, max 10 PDFs per call.<\/li>\n<\/ul>\nAvailability<\/h2>\n
\n
agents.defaults.pdfModel<\/code><\/li>\nagents.defaults.imageModel<\/code><\/li>\npdf<\/code> tool is not exposed.<\/p>\nInput reference<\/h2>\n
\n
pdf<\/code> (string<\/code>): one PDF path or URL<\/li>\npdfs<\/code> (string[]<\/code>): multiple PDF paths or URLs, up to 10 total<\/li>\nprompt<\/code> (string<\/code>): analysis prompt, default Analyze this PDF document.<\/code><\/li>\npages<\/code> (string<\/code>): page filter like 1-5<\/code> or 1,3,7-9<\/code><\/li>\nmodel<\/code> (string<\/code>): optional model override (provider\/model<\/code>)<\/li>\nmaxBytesMb<\/code> (number<\/code>): per-PDF size cap in MB<\/li>\n<\/ul>\n\n
pdf<\/code> and pdfs<\/code> are merged and deduplicated before loading.<\/li>\npages<\/code> is parsed as 1-based page numbers, deduped, sorted, and clamped to the configured max pages.<\/li>\nmaxBytesMb<\/code> defaults to agents.defaults.pdfMaxBytesMb<\/code> or 10<\/code>.<\/li>\n<\/ul>\nSupported PDF references<\/h2>\n
\n
~<\/code> expansion)<\/li>\nfile:\/\/<\/code> URL<\/li>\nhttp:\/\/<\/code> and https:\/\/<\/code> URL<\/li>\n<\/ul>\n\n
ftp:\/\/<\/code>) are rejected with unsupported_pdf_reference<\/code>.<\/li>\nhttp(s)<\/code> URLs are rejected.<\/li>\nExecution modes<\/h2>\n
Native provider mode<\/h3>\n
anthropic<\/code> and google<\/code>.
\nThe tool sends raw PDF bytes directly to provider APIs.<\/p>\n\n
pages<\/code> is not supported. If set, the tool returns an error.<\/li>\n<\/ul>\nExtraction fallback mode<\/h3>\n
\n
agents.defaults.pdfMaxPages<\/code>, default 20<\/code>).<\/li>\n200<\/code> chars, render selected pages to PNG images and include them.<\/li>\n\n
4,000,000<\/code>.<\/li>\npdfjs-dist<\/code> (and @napi-rs\/canvas<\/code> for image rendering).<\/li>\n<\/ul>\nConfig<\/h2>\n
\n{
\n agents: {
\n defaults: {
\n pdfModel: {
\n primary: “anthropic\/claude-opus-4-6”,
\n fallbacks: [“openai\/gpt-5-mini”],
\n },
\n pdfMaxBytesMb: 10,
\n pdfMaxPages: 20,
\n },
\n },
\n}<\/p>\n\nSee [Configuration Reference](\/gateway\/configuration-reference) for full field details.\n\n## Output details\n\nThe tool returns text in `content[0].text` and structured metadata in `details`.\n\nCommon `details` fields:\n\n* `model`: resolved model ref (`provider\/model`)\n* `native`: `true` for native provider mode, `false` for fallback\n* `attempts`: fallback attempts that failed before success\n\nPath fields:\n\n* single PDF input: `details.pdf`\n* multiple PDF inputs: `details.pdfs[]` with `pdf` entries\n* sandbox path rewrite metadata (when applicable): `rewrittenFrom`\n\n## Error behavior\n\n* Missing PDF input: throws `pdf required: provide a path or URL to a PDF document`\n* Too many PDFs: returns structured error in `details.error = "too_many_pdfs"`\n* Unsupported reference scheme: returns `details.error = "unsupported_pdf_reference"`\n* Native mode with `pages`: throws clear `pages is not supported with native PDF providers` error\n\n## Examples\n\nSingle PDF:\n\n```json theme={"theme":{"light":"min-light","dark":"min-dark"}}\n{\n "pdf": "\/tmp\/report.pdf",\n "prompt": "Summarize this report in 5 bullets"\n}\n<\/code><\/pre>\n
\n{
\n “pdfs”: [“\/tmp\/q1.pdf”, “\/tmp\/q2.pdf”],
\n “prompt”: “Compare risks and timeline changes across both documents”
\n}<\/p>\n\nPage-filtered fallback model:\n\n```json theme={"theme":{"light":"min-light","dark":"min-dark"}}\n{\n "pdf": "https:\/\/example.com\/report.pdf",\n "pages": "1-3,7",\n "model": "openai\/gpt-5-mini",\n "prompt": "Extract only customer-impacting incidents"\n}\n<\/code><\/pre>\n","protected":false},"excerpt":{"rendered":"