How It Works · Integration

However your content moves, Cascade moves with it.

API, print pipeline, e-reader, browser: Cascade fits the format you already use. No rebuild required.

Where it fits

One layer, however you reach your reader.

Pick the path that matches how your content already moves. Every path leads back to the same engine.

Real-time

API

Send text, get structure back in milliseconds. Built for any pipeline that can make an HTTP call.

Batch

Print & Export

Pre-process content once, output Cascade-formatted text for static and print-ready delivery.

Long-form

E-Readers & Reading Apps

Drop into long-form reading environments where comprehension depth matters most.

Live, in-page

Browser

The Cascade extension reformats any page a reader visits, on the fly, without touching the source site.

Drop-in

Direct Embed

A ready-made component for platforms that want Cascade’s presentation layer without building against the API.

Under the hood

Five steps between your text and a better reading experience.

You send text

Plain text, via the API, a CMS export, or a batch upload. Cascade doesn't care where it came from.

Cascade does this part automatically

Cascade parses the structure

NLP identifies clause boundaries, phrase groupings, and sentence shape — the linguistic analysis Cascade needs before it can act.

Cascade shapes the text

Cascade's algorithm decides where each line breaks and how each phrase is indented — turning linguistic structure into a reading pattern.

You get back data, not a picture

JSON, HTML, or plain text. Structure only — no visual styling is imposed.

You render it, your way

Your CSS, your product, your brand. Cascade never touches how it looks on screen or in print.

The API

Three ways in, one engine underneath.

Same request, three response shapes: pick whichever fits how your stack renders content.

POST /v1/cascade/json

Structured metadata: indent levels and line breaks only. Full control over how it renders.

POST /v1/cascade/html

Ready-to-render markup, styled with your own CSS classes. No client-side parsing needed.

POST /v1/cascade/text

Cascade-formatted plain text, for environments that can’t render HTML or parse JSON.

POST /v1/cascade/json

# response — structure only; your CSS does the styling
{
  "sentences": [
    {
      "json_cascade": [
        { "indent_level": 1, "line": "When the men" },
        { "indent_level": 1, "line": "hunt," },
        { "indent_level": 0, "line": "the birds" },
        { "indent_level": 1, "line": "with bright feathers" },
        { "indent_level": 0, "line": "typically scatter." }
      ]
    }
  ]
}

Auth

OAuth client credentials

Versioning

Semantic versioning: no breaking changes without a version bump

Input limit

32,000 characters per request; batch longer content across multiple calls

Errors

Standard HTTP status codes: 200 success, 401 unauthorized, 422 unprocessable entity, 500 server error

Security & ownership

Your content, your infrastructure, your rules.

Your content stays yours

Privacy-first by design. Deployment options keep your content inside your own systems. Cascade processes structure. It doesn’t store your text.

You control the rendering

Cascade returns structured output. Your CSS, your brand, your product. The presentation is yours to own. Cascade only shapes the underlying structure.

Your integration is closer than you think.

Request API access, or talk to us about the format that fits your stack: print, browser, e-reader, or direct embed.

Good to know

Common questions

How would we actually integrate Cascade into our platform or pipeline?

Cascade integrates via a simple API (endpoints for JSON, HTML, or plain text), a browser extension, or a direct platform embed, depending on your use case. See Integration for the full detail. Reach out to the Cascade team if you have questions about integrating Cascade with any of your products or systems.

What does the API return, exactly?

Three response options: JSON for structured metadata, HTML for ready-to-render markup styled with your own CSS classes, or plain text.

How is the API secured, and what happens if a request fails?

Authentication is handled through OAuth client credentials. The API returns standard HTTP status codes (200, 401, 422, 500) so failures are easy to detect and handle.

Is there a limit to how much content we can send through the API?

Yes. Requests are limited to 32,000 characters. In practice, submissions should be a sentence or paragraph length at most with multiple calls used to complete longer passages.

Does Cascade store or retain our content?

No. Cascade processes structure only and doesn’t store the underlying text. You retain control over your content.

What kinds of content and languages does it support?

Any content type, from news articles to legal contracts to scientific papers. It’s live today in English, French, German, Italian, and Spanish, with Mandarin coming soon.

How does Cascade interact with AI-generated content?

Cascade is a reading layer that benefits readers of all ages and levels, wherever they read. It makes AI-generated content, search results, and documentation easier to understand, not just easier to produce.

How do we get started or ask technical questions?

Request API access directly from the Integration page, or reach the team through the Contact page. It routes you to the right conversation, whether you’re evaluating Cascade for a team or a platform integration.

See all Cascade FAQs →