For AI agents: a documentation index is available at /llms.txt — every docs page also has a markdown twin (append .md to the page URL).
Integrations
Two ways to connect. The publishers below render every page of your docs site to PDF at build time; the agent-framework tools let an agent render a PDF on demand. Both wrap the same hosted makesPDF API.
Docs publishers
Pick the publisher that matches your static-site generator. All three render through the same POST /api/v1/md endpoint, so the PDFs look identical whichever one you use.
| Framework | Install | Guide |
|---|---|---|
| MkDocs | pip install "makespdf[mkdocs]" | MkDocs → PDF |
| Docusaurus | npm install --save-dev @makespdf/docusaurus-plugin | Docusaurus → PDF |
| Hugo | makespdf-pdf.sh (copy or curl) | Hugo → PDF |
What every publisher does
- One billed render per page. Each page is a single POST /api/v1/md call, billed at 1 credit per 10 pages. Combined mode adds one more render for the combined file.
- Per-page opt-out. Put pdf: false in a page’s front-matter to skip it entirely — no PDF, no bill.
- Combined mode. Optionally write one extra PDF with every markdown-rendered page concatenated, for a downloadable handbook next to the per-page files.
- Fail loudly. A failed render fails the build by default, so a deploy cannot silently ship without PDFs.
- No offline rendering. Every call is a hosted API call against https://makespdf.com, so billing matches the web app and CLI.
Agent frameworks
The python shims expose exactly two tools — markdown → PDF via POST /api/v1/md, and a stored template or inline DSL → PDF via POST /api/v1/render — and each writes a temp .pdf and returns its path. The Vercel AI SDK page defines the same two as tool() adapters. Every framework page renders the same shared demo data (MARKDOWN_DEMO, DSL_DEMO, DATA_DEMO, TEMPLATE_ID_DEMO), so the examples are directly comparable.
| Framework | Install | Guide |
|---|---|---|
| LangChain (Python) | pip install "makespdf[langchain]" | LangChain → PDF |
| LlamaIndex (Python) | pip install "makespdf[llamaindex]" | LlamaIndex → PDF |
| CrewAI (Python) | pip install "makespdf[crewai]" | CrewAI → PDF |
| LangGraph (Python) | pip install "makespdf[langchain]" langgraph | LangGraph → PDF |
| Vercel AI SDK (TypeScript) | npm install ai zod | Vercel AI SDK → PDF |
No-code automation
Automation platforms reach makesPDF through its hosted MCP server, so every MCP tool is available as an action.
| Platform | Connect with | Guide |
|---|---|---|
| Make | MCP Client app, makesPDF API key | Make → PDF |
Calling the API yourself
Each publisher wraps the same endpoint, so an integration you write yourself needs only an API key. For backend rendering and the hosted MCP server, see Integrate makesPDF into your app; the full Vercel AI SDK reference is on the Vercel AI SDK page. The endpoint contract is in the API reference.