PDF rendering API for agents and pipelines

Accessible, archive-grade PDFs from Markdown or template DSL — PDF/A-2A + PDF/UA-1 compliant, sub-second renders.

Validated against veraPDF 1.28.2 · 153 PDF/A-2A + 106 PDF/UA-1 rules · zero failures
Generate

Make a new PDF

Markdown, embedded AI chat, prebuilt template, or VS Code / Obsidian / MCP. Same compliant pipeline behind all four.

Enhance

Remediate an existing PDF

PDF/UA-1 accessibility tags + PDF/A-2A archival on a legacy PDF. Tagged original plus a clean DSL source you own — for EAA, ADA Title II, Section 508 compliance.

Validate

Check a PDF

Hosted veraPDF 1.28.2 — PDF/A-2A and PDF/UA-1 in one upload. Free, no sign-in, no retention.

Built for agents

Skill, MCP, CLI, x402

One skill file teaches Claude, Cursor, ChatGPT, or any agent to author templates. Native MCP server. Pay-per-render via x402 wallets — no account required.
Compliant by default

PDF/A-2A + PDF/UA-1

Tagged structure tree, embedded fonts, Unicode text extraction, sRGB ICC profile. Validated against veraPDF every release.
Markdown-fluent

GFM, mermaid, math, alerts

GitHub Flavored Markdown in, styled PDF out. Tables, code blocks, task lists, Mermaid diagrams, LaTeX math, GFM alerts — all native PDF vector graphics.

Works withClaude Code·Cursor·ChatGPT·Windsurf·Claude Design·any MCP host

See how we compare to react-pdf, Puppeteer, and 8 more →

Tell your AI agentPaste into Claude, Cursor, ChatGPT, etc.
Use this skill to author PDFs: https://makespdf.com/skills/pdf-template-author.md
Author with our embedded AI
Two-column chat at /ai/chat: describe a layout, see the PDF rebuild in the workspace pane, save the result as a reusable template. No model of your own required.
…or bring your own
One skill file teaches Claude, Cursor, ChatGPT, or any agent to author templates against our DSL. MCP server too. No lock-in.
Fast renders
Deterministic DSL pipeline. Same template, new data, sub-second per PDF. No AI in the hot path.
Draft → publish
Iterate free on /preview. Save the template. Render with data on /render — billed, clean, production-grade.
Markdown to PDF
GitHub Flavored Markdown in, styled PDF out. Tables, code blocks, task lists.
API-First
Built for agents and automation. REST API with JSON in, PDF bytes out.
PDF/A-2A + PDF/UA-1
Dual-compliant archival and accessible PDFs. Full structure tree, embedded fonts, Unicode text extraction.
Fillable forms
Real AcroForm widgets — text fields, checkboxes, dropdowns — that stay interactive in Reader and Preview, and are tagged for PDF/UA-1 screen readers.

Markdown to PDF

Paste or upload Markdown and get a styled PDF instantly

VS Code extension

Convert Markdown to PDF without leaving your editor — command, title-bar icon, and right-click action

Core API endpoints
EndpointWhat it doesBilling
POST /api/v1/mdConvert GitHub Flavored Markdown to a styled, accessible PDF — sub-second for typical docs.Billed
POST /api/v1/previewDraft and iterate on a template with inline DSL — the authoring endpoint.Free
POST /api/v1/renderRender a saved or inline template with data — the production endpoint, same deterministic pipeline as preview.Billed
# Render a PDF from a template
curl -X POST https://makespdf.com/api/v1/render \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"templateId": "tmpl_abc", "data": {"company": "Acme"}}' \
  -o invoice.pdf

Embedded chat · Bring your own AI · Skill file · llms.txt

How does pricing work?

The web UI and template authoring are free. API renders are billed at 1 credit per 10 pages — the free tier includes 50 credits per month. You can also pay per-render in USDC via x402 wallets with no account required. See the pricing page for plan details and credit top-ups.