extracting-document-text
Skill0 starsChanged 9 days agoMITView on GitHub
Pulls text out of documents and images with upAPI: extracts a PDF's text (by URL or upload, page ranges, tables as Markdown), reads the text in an image with OCR (lines with confidence and position), and counts characters, words, lines and sentences with a SHA-256 digest. Use when the user asks 'get the text out of this PDF', 'what does page 3 of this report say', 'extract the tables from this PDF', 'read this receipt', 'OCR this screenshot', or 'how many words is this'. Read-only; spends upAPI quota.
In this repository66 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- adding-uptime-monitorsSkill
- analyzing-page-performanceSkill
- booking-meetingsSkill
- building-onboarding-workflowsSkill
- building-study-guidesSkill
- capturing-tasksSkill
- capturing-web-pagesSkill
- chasing-unpaid-invoicesSkill
- checking-domain-connectionsSkill
- checking-email-deliverabilitySkill
- checking-notification-deliverySkill
- checking-uptime-statusSkill
- cleaning-up-bookkeepingSkill
- connecting-customer-domainsSkill
- discussing-buildsSkill
- drafting-invoicesSkill
- drafting-social-postsSkill
- editing-demosSkill
- editing-page-draftsSkill
- finding-meeting-timesSkill
- handling-flaky-changesSkill
- investigating-tasksSkill
- making-demo-videosSkill
- managing-bookingsSkill
- managing-content-calendarSkill
- managing-demosSkill
- managing-voice-librarySkill
- planning-revisionSkill
- preflighting-domainsSkill
- publishing-pagesSkill
- reading-financial-reportsSkill
- reading-text-aloudSkill
- reading-web-pagesSkill
- reporting-snapvisor-usageSkill
- researching-course-materialSkill
- researching-summary-librarySkill
- responding-to-incidentsSkill
- reviewing-email-performanceSkill
- reviewing-notification-workflowsSkill
- reviewing-social-performanceSkill
- reviewing-visual-buildsSkill
- reviewing-workloadSkill
- searching-getitdone-api-docsSkill
- sending-email-campaignsSkill
- sending-notificationsSkill
- setup-bioflowSkill
- setup-calySkill
- setup-demofySkill
- setup-dodomainSkill
- setup-getitdoneSkill
- setup-notiflySkill
- setup-postifySkill
- setup-sendlySkill
- setup-shortySkill
- setup-snapvisorSkill
- setup-superbooksSkill
- setup-unotesSkill
- setup-upapiSkill
- setup-uptimelySkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Every tool in this skill comes from the upapi MCP server, the upAPI connector, and is written here as upapi:<tool>.
- Very specific: line 15
Tools listed by name (the plugin's
?tools=claudesurface): call them directly, for example upapi:pdf_extract_text_post.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
<!-- Generated from plugins/upapi/skills/extracting-document-text/SKILL.md by scripts/sync-skills.mjs. Edit the plugin copy, then run the script. -->upAPI: extracting document text
Turn a PDF or an image into text, then answer from that text.
Connecting and calling the tools
Every tool in this skill comes from the upapi MCP server, the upAPI connector, and is written here as upapi:<tool>.
- Tools listed by name (the plugin's
?tools=claudesurface): call them directly, for exampleupapi:pdf_extract_text_post. - Only
upapi:search_opsandupapi:call_oplisted (the plain URL's compact table): callupapi:call_opwith the slug and input, for example{ "slug": "pdf-extract-text.post", "input": { "url": "https://example.com/report.pdf" } }. - No
upapitools at all: upAPI is not connected yet. Follow thesetup-upapiskill.
A failed operation comes back as an error result whose text starts with its code, for example INVALID_INPUT:.
Tools you will use
upapi:pdf_extract_text_post(slugpdf-extract-text.post, 2 units): extracts a PDF's text. Input:urlof a public PDF orfileas base64 (nodata:prefix, at most 3.7 MB decoded), not both;pages(1-based, for example1-5,8);tables(default false; slower, returns each detected table as Markdown).upapi:image_ocr_post(slugimage-ocr.post, 4 units): reads the text in an image. Input:urlorfileas base64 (at most 3.7 MB decoded), not both;languages(enandchare supported, one model covers Latin and Chinese). Returns the full text plus every line with its confidence and bounding box.upapi:text_analyze_post(slugtext-analyze.post, 1 unit): counts characters, words, lines, sentences and bytes of atext, and returns its SHA-256 digest. Runs without network access.
Workflow
- Find the source. A public URL is best. A local file works in Claude Code: read it, base64-encode it, and send it as
fileif it is under 3.7 MB; for anything larger ask for a URL. - For a PDF, set
pageswhen the user asks about part of it, and settables: trueonly when they want tables. - For an image (a receipt, a scanned page, a screenshot of text), call
upapi:image_ocr_post. Mention low-confidence lines when they matter to the answer. - Answer from the extracted text: quote, summarize, or pull out the fields the user asked for. Say which pages or lines you used.
- For "how long is this" questions, call
upapi:text_analyze_poston the text and report its counts rather than estimating.
Rules
- These tools are for the user's own documents and media. Do not run them on ID documents, medical records, or payment-card details. upAPI does not keep the file after the call; the extracted text may sit in its response cache for up to 24 hours.
- A scanned PDF with no text layer returns little or no text; say so and offer OCR on page images instead of guessing the content.
- On
RATE_LIMITED, wait the seconds the message gives before one retry. OnQUOTA_EXCEEDED, stop and tell the user.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

