Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
AI agents generate PDFs by calling a controlled application tool—not by producing text alone. A practical workflow is: define the report, validate the agent’s structured output, render semantic HTML with CSS in a browser (or draw PDF primitives directly), save the file, and return it through an application-controlled storage or download path. Puppeteer documents Page.pdf() for this HTML-to-PDF route; Playwright also exports PDFs, but its documented PDF feature is Chromium-only.
The complete agent-to-PDF workflow
Separate reasoning from execution. The model can plan sections, summarize approved data and fill a report schema, but application code should decide whether a file may be created, which inputs are allowed and where the result is written.
- Define the deliverable. Specify the title, audience, page size, language, required sections, data sources and output format (PDF, filename and retention period).
- Expose a narrow tool. Give the agent a function such as
create_reportwith typed fields for content and rendering options. Do not expose an unrestricted shell, filesystem or network client when a purpose-built function is sufficient. - Validate in application code. Enforce length limits, allowed URLs, CSS options, page ranges and output paths. Reject unknown fields and malformed values before a renderer starts.
- Build the document. Convert the validated result into semantic HTML and CSS for browser printing, or construct pages with a PDF library when direct drawing is a better fit.
- Render and verify. Launch the approved renderer, wait for required fonts or assets, create the PDF and check that a non-empty file was produced.
- Deliver through your boundary. Store the artifact in an approved location, return a short-lived download reference, or attach it to your application response. The delivery mechanism is application-specific.
OpenAI’s tools documentation describes function tools, MCP connections and sandbox configuration as ways to provide capabilities. An agent response by itself is not a PDF-generation mechanism.
Recommended Free Tools
Choose HTML-to-PDF or direct PDF construction
HTML and CSS rendered by a browser
Use this route when your report already exists as a web view or needs CSS layout, tables, responsive components and familiar print rules. Your application produces HTML, then a browser prints that page. Puppeteer’s PDF guide documents Page.pdf() and says it waits for fonts to load by default.
#1 Best Overall
- Scanner type: Document
- Connectivity technology: USB
- With Auto Scan Mode, the scanner automatically detects what you're scanning
- Digitize documents and images
Direct PDF-library output
Use drawing primitives when you need explicit coordinates, predictable pagination or a library-specific PDF object model. The agent still should return structured content; trusted application code converts that content into pages. Do not let model-generated code choose arbitrary files or network destinations.
Decision table
| Question | HTML-to-browser | Direct PDF construction |
|---|---|---|
| Input representation | HTML and CSS | PDF-library elements or drawing commands |
| Runtime requirement | Browser executable, fonts and assets | Chosen PDF library and its runtime |
| Best fit | Existing web layouts, styled reports and tables | Precise coordinates and custom pagination |
| Operational concerns | Browser startup, asset loading and print CSS | Font embedding, wrapping and page-break logic |
| Output handling | Application validates, stores and presents the resulting file | |
Neither cited browser guide establishes that one renderer is universally more reliable or produces better PDFs. Select based on your layout and operational constraints.
Build a safe report-generation tool
A constrained schema
Keep the tool contract small. A useful schema might include title, an array of sections with plain text or approved HTML fragments, paperSize limited to known values, landscape as a boolean, and a server-selected output identifier. Exclude arbitrary shell commands, unrestricted JavaScript and caller-supplied absolute paths.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchValidate content and assets
- Set maximum lengths for titles, paragraphs, tables and total document size.
- Allow only tags and attributes your sanitizer understands if the agent supplies HTML.
- Permit images and styles only from approved locations; reject internal IP ranges when fetching remote assets.
- Normalize filenames and create a server-owned temporary directory for each job.
- Return structured errors such as
INVALID_INPUT,RENDERER_UNAVAILABLEandDELIVERY_FAILED.
Use a runtime only when needed
The Agents API quickstart demonstrates an agent writing and running a script in an OpenAI-hosted sandbox. It also notes that an environment of none is appropriate when a task does not need code execution or local files. If your workflow only drafts report text for a human, do not grant a code-execution environment; if it must render and manipulate a file, provide the smallest isolated runtime that can do so.
Rank #2
- PORTABLE SCANNER FOR USE ON-THE-GO — The fastest and lightest mobile single-sheet-fed compact document scanner in its class¹
- QUICK DOCUMENT SCANNING ― This Epson ultra-fast scanner scans a single page as quickly as 5.5 seconds²; Windows and Mac compatible
- VERSATILE PAPER HANDLING ― Portable scanner scans documents up to 8.5 x 72 in; Also easily digitizes receipts and ID cards to make accounting, bookkeeping, and organizing simpler
- INTUITIVE, HIGH-SPEED SOFTWARE — Epson ScanSmart Software³ is a smart tool allowing you to easily scan, review, and save; Stay organized easily with the help of this Epson scanner
- EASY SETUP — USB-powered connect to your computer for quick and simple scanning; No batteries or external power supply required to operate portable document scanner; Standard Connectivity: USB 2.0
Runnable Node.js example with Puppeteer
The following example accepts already-validated report data, creates semantic HTML, and prints it to a PDF. Install Puppeteer in your application, keep the generated HTML under your control, and choose an output directory owned by the service.
- Create a project and install Puppeteer:
npm install puppeteer. - Save this as
make-report.mjs. - Run
node make-report.mjs; it writesreport.pdf.
import puppeteer from 'puppeteer';
import { writeFile } from 'node:fs/promises';
const report = {
title: 'Quarterly operations report',
sections: [
{ heading: 'Summary', body: 'Revenue remained stable while support response time improved.' },
{ heading: 'Next steps', body: 'Document the approved actions and owners before distribution.' }
]
};
function escapeHtml(value) {
return String(value).replace(/[<>&"']/g, ch => ({
'&': '&', '<': '<', '>': '>', '"': '"', "'": '''
}[ch]));
}
const html = `<!doctype html>
<html><head><meta charset="utf-8">
<style>
@page { size: A4; margin: 18mm; }
body { font-family: Arial, sans-serif; color: #202124; line-height: 1.45; }
h1 { font-size: 24px; margin: 0 0 16px; }
h2 { font-size: 16px; margin: 18px 0 6px; }
</style></head><body>
<h1>${escapeHtml(report.title)}</h1>
${report.sections.map(s => `<h2>${escapeHtml(s.heading)}</h2><p>${escapeHtml(s.body)}</p>`).join('')}
</body></html>`;
const browser = await puppeteer.launch({ headless: true });
try {
const page = await browser.newPage();
await page.setContent(html, { waitUntil: 'networkidle0' });
await page.pdf({ path: 'report.pdf', format: 'A4', printBackground: true });
} finally {
await browser.close();
}
For production, add a job timeout, a maximum HTML size, a cleanup policy for temporary files and a post-render check that the PDF exists and is non-zero. If your page depends on remote images, make the network allowlist explicit rather than allowing every destination.
Playwright and browser-specific constraints
Playwright’s PDF export can save the current page as a PDF, but its documentation specifies that PDF generation is Chromium-only. If your deployment selects Firefox or WebKit, use Chromium for this step or choose another rendering path. Keep the browser version and installed fonts consistent across environments so pagination does not change unexpectedly.
Security: treat the agent and its inputs as untrusted
OpenAI’s sandbox security guidance states: “Agent-generated code can access the files, credentials, and network available to its environment.” Isolate each workload, restrict outbound traffic to approved endpoints and separate application credentials from the environment. Keep API keys out of generated code and agent-visible files; use a secrets manager or a trusted proxy for third-party requests.
Rank #3
- FAST SPEEDS - Scans color and black and white documents a blazing speed up to 16ppm (1). Color scanning won’t slow you down as the color scan speed is the same as the black and white scan speed.
- ULTRA COMPACT – At less than 1 foot in length and only about 1. 5lbs in weight you can fit this device virtually anywhere (a bag, a purse, even a pocket).
- READY WHENEVER YOU ARE – The DS-640 mobile scanner is powered via an included micro USB 3. 0 cable allowing you to use it even where there is no outlet available. Plug it into you PC or laptop and you are ready to scan.
- WORKS YOUR WAY – Use the Brother free iPrint&Scan desktop app for scanning to multiple “Scan-to” destinations like PC, Network, cloud services, Email and OCR. (2) Supports Windows, Mac and Linux and TWAIN/WIA for PC/ICA for Mac/SANE drivers. (3)
- OPTIMIZE IMAGES AND TEXT – Automatic color detection/adjustment, image rotation (PC only), bleed through prevention/background removal, text enhancement, color drop to enhance scans. Software suite includes document management and OCR software. (4)
Reports may contain uploaded files, retrieved web pages or user text. Treat all of that as data, not instructions. Prompt injection can attempt to change tool behavior, while private-data leakage can expose content in the PDF or in logs. OpenAI’s agent-safety guidance recommends structured outputs, clear instructions, input guardrails, human approval for MCP operations and evaluation of agent traces. These controls reduce risk; they do not guarantee correctness.
- Require approval before sending a consequential or sensitive report.
- Redact secrets and personal data before rendering when they are not required.
- Log tool inputs, validation outcomes and artifact identifiers, but not raw confidential content unless policy permits it.
- Use separate service accounts for rendering and storage, with the minimum permissions.
Reliability, performance and cost considerations
Fonts and layout
Missing fonts, unsupported glyphs, overflowing tables and accidental page breaks are common causes of unusable PDFs. Package required fonts with the renderer, use print-specific CSS, and test long headings, wide tables, images and right-to-left or non-Latin text if your users need them.
Network and assets
Every remote stylesheet, image or web font adds a failure point. Prefer local, versioned assets; set finite navigation and resource timeouts; and fail clearly when a required asset cannot load. Do not silently substitute an empty image or blank page.
Free tools Windows power users keep installed
One-click scans. No signup required.
Throughput
Launching a browser for every request is simple but expensive in latency. A controlled worker pool can reuse browser processes while creating a fresh page per job. Cap concurrent pages, enforce per-job time limits and recycle unhealthy workers. Measure your own workload rather than applying a generic benchmark.
Rank #4
- FAST DOCUMENT SCANNING — Document scanner with feeder allows you to speed through stacks with a 50-sheet Auto Document Feeder (ADF); Efficient office scanner to help you scan more productively
- INTUITIVE, HIGH-SPEED SOFTWARE — Quickly scan with this desktop document scanner; Epson ScanSmart Software lets you easily preview scans, email files, upload to the cloud, and more; Plus, automatic file naming saves even more time
- SEAMLESS INTEGRATION — Easily incorporate your data into most document management software with the included TWAIN driver; Office document scanner integrates seamlessly with business workflows
- EASY SHARING — Duplex scanner allows you to scan straight to email or popular cloud storage2 services like Dropbox, Evernote, Google Drive, and OneDrive for simple storage and sharing
- SIMPLE FILE MANAGEMENT — Scanner allows the creation of searchable PDFs with Optical Character Recognition (OCR) and convert scans to editable Word or Excel files effortlessly; Designed for home and office document scanning
Cost model
Budget for model calls, agent runtime, browser or PDF-library compute, storage and outbound transfer. Retries should be limited and idempotent: use a job identifier so a retry does not publish duplicate files. Cache immutable inputs where policy allows, but never reuse a PDF when its source data or permissions changed.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Troubleshooting common failures
| Symptom | Likely cause | Fix |
|---|---|---|
| Browser fails to start | Missing executable, sandbox restriction or incompatible container | Install the supported browser dependencies, verify the executable path and run the renderer in an isolated worker with the permissions it needs. |
| Blank or nearly empty PDF | HTML was set before content loaded, or a script replaced the page | Wait for the required selector or network state, then capture; record the final URL and page errors. |
| Fonts or icons are missing | Font files were blocked, unavailable or not embedded | Bundle approved fonts, allow their paths, and wait for font loading before printing. |
| Images are absent | Remote request failed, URL was blocked or lazy loading never ran | Use approved asset URLs, wait for the image selector, and load lazy content before capture. |
| Tables split badly | Print CSS lacks break rules or cells are too wide | Set print widths, add page-break rules and provide a landscape variant for genuinely wide tables. |
| Agent changes the requested format | Untrusted document text influenced tool instructions | Keep tool instructions outside document content, validate the structured result and require approval for format changes. |
| File cannot be downloaded | Storage or delivery step failed after rendering | Keep the artifact until delivery is confirmed, return a typed error and make cleanup retry-safe. |
Or skip the browser setup
ScreenshotNeo is a website screenshot API and MCP server that can capture a rendered page as PNG, JPEG, WebP or PDF. It accepts consent banners before capture and removes more than 60 known consent platforms, newsletter popups and chat widgets; each step can be disabled. Bot checks, blank pages, timeouts, failed loads and cache hits are not billed, and the response identifies the page verdict and billing result in X-Page-Verdict and X-Billed headers. Its MCP server provides take_screenshot, get_page_info and capture_pdf tools for Claude, Cursor and other MCP clients.
One call can return a PDF of a public report page:
curl -G "https://api.screenshotneo.com/v1/shot" -d access_key=YOUR_API_KEY --data-urlencode url=https://stripe.com -o shot.webpFor PDF output, set the service’s PDF options in the request; the API also supports paper size, margins, landscape mode and page ranges. See the ScreenshotNeo documentation for parameter details.
Python:
import requests
r = requests.get("https://api.screenshotneo.com/v1/shot", params={"access_key": "YOUR_API_KEY", "url": "https://stripe.com"}, timeout=90)
r.raise_for_status()
open("shot.webp", "wb").write(r.content)Node.js:
const q = new URLSearchParams({ access_key: 'YOUR_API_KEY', url: 'https://stripe.com' });
const res = await fetch(`https://api.screenshotneo.com/v1/shot?${q}`);
if (!res.ok) throw new Error(`HTTP ${res.status}`);
const bytes = new Uint8Array(await res.arrayBuffer());
await Bun.write('shot.webp', bytes);The Free plan includes 1,000 shots per month with no card; paid plans start at $5 for 3,000 shots. Create a free ScreenshotNeo account to try the API.
FAQ
Does an AI model create the PDF by itself?
No. The model supplies content or a plan; a configured tool and runtime perform rendering and file handling.
Best Value
- FITS SMALL SPACES AND STAYS OUT OF THE WAY. Innovative space-saving design to free up desk space, even when it's being used
- SCAN DOCUMENTS, PHOTOS, CARDS, AND MORE. Handles most document types, including thick items and plastic cards. Exclusive QUICK MENU lets you quickly scan-drag-drop to your favorite computer apps
- GREAT IMAGES EVERY TIME, NO EXPERIENCE REQUIRED. A single touch starts fast, up to 30ppm duplex scanning with automatic de-skew, color optimization, and blank page removal for outstanding results without driver setup
- SCAN WHERE YOU WANT, WHEN YOU WANT. Connect with USB or Wi-Fi. Send to Mac, PC, mobile devices, and cloud services. Scan to Chromebook using the mobile app. Can be used without a computer
- PHOTO AND DOCUMENT ORGANIZATION MADE EFFORTLESS. ScanSnap Home all-in-one software brings together all your favorite functions. Easily manage, edit, and use scanned data from documents, receipts, business cards, photos, and more
When should I require human approval?
Require it before distributing reports that contain regulated, private or consequential information, or when an agent requests a permission or destination outside its normal schema.
Can I use Playwright without Chromium?
Not for the documented PDF export feature; Playwright specifies Chromium-only PDF generation.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Where should the generated file be stored?
Use an application-controlled storage layer with scoped permissions, retention rules and an authenticated delivery path rather than an agent-selected location.
Frequently Asked Questions
Can an agent generate a PDF without code execution?
It can draft content, but creating and saving a PDF requires application code, a connected tool or another renderer.
What should I monitor in production?
Track validation failures, render duration, timeout rates, missing assets, output size and delivery success, while protecting sensitive report content in logs.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →
