- Home
- Skills
- Productivity
- PDF to Word Studio
More screenshots
Works with the AI Agents you already use
PDF to Word Studio
Rebuild PDFs, scans and document photos as refined, genuinely editable Word files, with source-linked checks and 36 practical scenarios.
Free
About this skill
Your PDF. Ready to edit.
Turn an existing PDF, scan or document photograph into a refined Word document you can keep working on. PDF to Word Studio is a complete free English Skill: no paid upgrade, required seller backend or paid OCR API.
Preserve the content, then improve the page
Reconstruct native Word paragraphs, headings, table cells and separately replaceable pictures. Keep names, identifiers, dates, amounts, negative signs, meaningful blanks and literal zeroes. Review source-page coverage and exact table associations. Unclear or unsupported regions must be identified and resolved before a complete handoff is claimed. No universal accuracy percentage or lossless-conversion guarantee is offered.
Choose layout-oriented reconstruction to retain source page size, clear columns and structure approximately. Choose clean reflow for coherent paragraphs, deliberate typography, comfortable spacing and consistent tables. Both routes retain source-page boundaries; neither promises a pixel-identical replica or automatic cross-page editorial redesign. Twelve document foundations support different purposes, from editorial reports and correspondence to ledgers, study materials and operational records.
36 complete source scenarios
Six families cover business briefs, research summaries, proposals, reviews, guidelines and release notes; handouts, study guides, research digests, workbooks, reading notes and lesson plans; quotations, invoices, price lists, stock records, expenses and landscape comparison matrices; resumes, creative resumes, letters, application checklists, case studies and references; intake forms, meeting records, itineraries, household checklists, rosters and run sheets; image-only letters, forms, quotations, document photos, mixed native/scanned PDFs and notices.
The six gallery boards show actual creator-local source pages and actual rendered DOCX results, with all 36 cases represented. The reusable package includes English scenario guidance and original fictional source recipes. Binary demo PDFs, finished demo DOCX files and the local gallery application are creator verification artifacts, not a promised downloadable demo bundle through Agensi MCP. Example content is never substituted for reading your document.
A practical revision workflow
Supply a document you are authorised to process, choose the pages and explain the intended edits. The Agent routes each page through its text layer or local English OCR, reconstructs the editable copy, checks native content and renders the result for review. Low-confidence OCR is highlighted. Regular ruled scan tables can become native Word tables; irregular tables, handwriting, damaged scans, equations and meaningful vector graphics may need manual reconstruction.
With the required Agent tools, generated work consists of the DOCX, a source-linked review record, an editable content map, recovered source pictures and concise editing notes. Source files remain unchanged. A source-map revision can be recompiled; separate manual Word edits do not automatically synchronise back to that map or the original PDF. Recovered figures remain pictures rather than editable chart data.
Requirements and verified scope
Requires a buyer Agent with Python 3.10+, a usable interpreter, file access and an authorised output location. The helper uses python-docx, pdfplumber, pypdf, Pillow and NumPy. English OCR needs local Tesseract English data; PDF scan rendering needs Poppler. Final layout review needs an available document renderer. Tool installation, compute and model usage depend on your Agent environment. No third-party binaries or fonts are redistributed.
Locally verified: 36 actual fictional source documents, 72 native DOCX outputs, 78 visually reviewed Word pages and 22 forward checks. Ordered source-reference content, exact critical currency and native table-cell associations passed for this fixture set. A meaningful native text/table edit was saved and reopened programmatically. Previews use headless LibreOffice.
Microsoft Word desktop edit/save, external customer documents, Agensi buyer-side multi-file/helper execution, dependency access and generated-file retrieval have not been verified. Loading SKILL.md alone does not establish that the conversion workflow works. Please have Agensi verify those execution and delivery paths before approval. Protected Skill instructions/helpers load through Agensi; generated customer work is a separate output and requires the buyer Agent's working file-delivery capability.
Questions
Is the complete Skill free? Yes. There is no paid feature gate. Your own Agent and computing environment may have costs.
Is the Word content editable? Essential recovered text and regular table cells are native Word objects. Separate source figures remain pictures. A full-page scan with invisible OCR text is not treated as editable reconstruction.
Can it read scans and photos? It supports local English OCR, mixed PDF pages and small-angle photo deskew. Unclear recognition requires source comparison; OCR confidence is not an accuracy certificate.
Will the layout match exactly? Fonts and geometry are reconstructed approximately. Reflow intentionally improves native typography while preserving content and source-page boundaries.
What about amounts, blanks and zeroes? They are checked separately, including cell placement, signs, decimals and leading zeroes. Source totals are retained rather than silently recalculated.
Are my files uploaded to an OCR service? The bundled helper uses local extraction/OCR and does not upload documents or follow embedded links. Your chosen Agent's own data handling still applies.
Does it work in every Agent? No. File access, dependencies, rendering and output retrieval must be available. Agensi buyer execution and Microsoft Word desktop acceptance remain unverified.
Original code and instructions are MIT licensed. Third-party dependencies retain their own licences. You must have the necessary rights to the source files you submit. The Skill is not affiliated with Microsoft or Adobe.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Productivity.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
4 uses
Times used to date
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 27 days ago
- Free with an Agensi account
- Software
Python