V3.0 — PRODUCTION ENGINE

The Language for AI Documents

Stop forcing LLMs to write HTML. DARE is a deterministic, token-efficient language with 18+ built-in components that generate pixel-perfect PDFs from simple prompts.

@setup { format: A4; // Reusable styles $brand: color=#2563eb; } @doc { cover { box(h=fill, p=20mm, bg=#1e293b) { txt(bold, color=white, size=28) { DARE v3 COVER } } } section { box(p=10mm) { badge(bg=#22c55e) { LIVE } math(expr="E = mc^2") {} } } }
0 Components
0 Less Tokens
0 Render Time
0 Deterministic

Why Not HTML?

HTML was built for fluid screens. DARE is built for fixed paper. It eliminates the chaos of CSS when generating PDFs via AI.

Token Efficient

DARE's compressed syntax uses 70% fewer tokens than HTML. Save money on every LLM API call with tags like box, txt, and bar.

Physical Units

No more pixels. DARE uses millimeters (mm) for all layout. If you write h=50mm, it prints exactly 50mm on paper.

Professional Charts

Built-in bar and pie chart components render vector-perfect SVG graphics directly from data strings.

Smart Tables

The tbl component creates professional grid tables from CSV-like data with auto-headers and zebra striping.

Header + Footer

Native hdr and ftr components with h=fill body expansion create perfect page layouts every time.

AI-Native Design

Ships with a ready-to-use System Prompt that teaches any LLM to write perfect DARE code. Copy, paste, generate.

What's New in V3

The DARE v3 production upgrade introduces cover page layouts, automated section breaking, math/chemistry structures, and direct download APIs.

COVER PAGES

Full-Bleed Cover Pages

Designed a two-pass rendering engine that compiles cover pages with zero margin/padding and no headers/footers, merging them with margins-enabled content pages in a single compiled PDF.

PAGINATION

Automated Section Breaks

No more manual page splitting. Use section { ... } tags to group flowable layout blocks. The compiler automatically splits sections across pages cleanly without text overlapping.

SCIENCE

KaTeX Math & Chemistry

Added native rendering for mathematical equations with the math component and chemical formula strings with the chem component.

API TOOLING

Direct Download Link API

Created GET request endpoints for our compiler API. AI Agents can now write DARE code and directly output a clickable download link for users to download their PDF documents instantly.

The Code Wars

The same dashboard header. One language was built for the web. The other was built for documents.

Standard HTML/CSS TOKENS: 0
DARE Engine v3 TOKENS: 0
73% FEWER TOKENS

Same visual result. Fraction of the cost for every AI API call.

Integration Workflows

DARE v3 is built for dual workflows. High-performance developer integrations meet token-efficient LLM rendering.

🤖 AI AGENTS

Gemini, Claude & ChatGPT

Provide DARE system instructions directly to your agent. Let the LLM design your layout and generate direct, clickable PDF download links in chat.

1
Configure Agent Prompt:

Provide DARE syntax guidelines in your assistant instructions. The agent outputs light markup instead of heavy HTML/CSS.

2
Direct Render GET API:

The agent converts the generated code into a URL-safe markdown download link using our REST API: https://dare.page.dev/api/render?code=[URL_ENCODED_DARE]

3
Immediate PDF Outputs:

Clicking the link renders the PDF instantly on our serverless engine, delivering the finished document with zero configuration.

"Create a resume from my Linkedin profile info"
🤖
Here is your custom formatted resume: Download Resume.pdf
💻 DEVS / NORMAL USERS

Local Compilation & Backend

Integrate DARE directly into your Node.js backends or compile documents using CLI commands during automated scripts.

1
CLI Compiler tool:

Use the local compiler CLI in shell tasks or CI/CD pipelines to build documentation: node cli.js input.dare output.pdf

2
Backend POST API:

Host the compiler on your own infrastructure and send rendering requests using raw POST bodies:

curl -X POST http://localhost:3000/api/render \
     --data-binary "@report.dare" \
     -o output.pdf
3
NodeJS Integration:

Import the compiler module directly into your codebase and handle buffers directly:

const dare = require('./src/parser');
const html = await dare.compile(code);
terminal

$ node cli.js input.dare report.pdf

Compiled successfully in 142ms!