Updates

PDF generation is solved. Every language has libraries. Every framework has integrations. So why build another PDF tool?

Because the problem isn't generation. It's customization.

When a customer asks to move a logo, change a font, or add a field, that request goes to a developer. They dig into template code, make changes, test across edge cases, deploy. A 5-minute design decision becomes a 2-hour engineering task.

Glyph changes this. Click any section of a PDF. Describe what you want in plain English. Watch it happen. No code. No deployments. No waiting.

We built Glyph because PDF customization should be as easy as asking for it.

"Add a watermark." "Make the header blue." "Put the QR code in the top right."

These are instructions anyone can give. Until now, only developers could execute them.

Glyph bridges that gap with AI. Select a region. Type what you want. The AI modifies the underlying HTML, validates the change, and renders a new preview instantly.

No template language to learn. No variables to configure. Just describe what you want.

Your invoice data is in Airtable. Your customer records are in Airtable. Your inventory is in Airtable.

Why should generating a PDF require a separate data pipeline?

Connect your Airtable base. Select a table. Glyph auto-detects your schema and maps fields to template regions. glyph.preview() and glyph.generate(). That's it.

Two API calls. One PDF. Zero data transformation.

Claude Code, Cursor, Windsurf. AI coding assistants are everywhere. But ask them to generate a PDF and you get suggestions, not documents.

Glyph's MCP server gives AI assistants actual PDF capabilities. "Create an invoice for this order" becomes a real invoice. "Generate a report from this data" becomes a downloadable PDF.

Install the server. Add your API key. Your AI assistant now speaks PDF.

This is what AI-native development looks like: tools that let AI do real work, not just write code about work.