Map the consequences before you commit.

Name: Cascade
Brand: Mariete
Availability: PreOrder

Describe a decision, an event, or a policy change. Cascade brings in six experts — an economist, a sociologist, a technologist, a political analyst, a lawyer and an environmental scientist — and shows you what each one thinks will happen next. First-order effects, then the effects of those, then the effects of those. Every link is explained and sourced.

Run a cascade See an example tree

Already running across 600+ teams on the Mariete platform.

cascade · overview

use case·Pricing

"What happens if we triple our prices?"

6specialist agents

What it does

Built around one idea: finish the work.

Configure once, run continuously. Every capability below is a production-grade surface, not a demo.

Six lenses, not one.

Economic, social, technological, political, legal, environmental. Each agent reasons within its domain; the tree composes their outputs into a coherent map.

First, second, third order.

Branches fork. A first-order consequence produces second-order effects, which produce third-order ones. Zoom, pan, collapse — the depth is navigable.

Every node is reasoned.

Click a node: agent reasoning, source citations, confidence. No black boxes — the tree is auditable all the way down.

Compare two scenarios side by side.

Run "status quo" and "proposed change" against the same prompt. Diff the trees; see what each decision uniquely causes.

In the wild

Five ways teams actually ship with Cascade.

Scenario · 01

"What happens if we triple our prices?"

Cascade maps more than 40 knock-on effects: revenue per customer goes up, but the cost to acquire new ones rises 40%, mid-market customers quietly leave in months 2–4, competitors undercut you, and three enterprise contracts flag potential breach risks.

Try this scenario

cascade · scenario · pricing

Intelligence suite

Four surfaces. One runtime. Zero duct tape.

Every Cascade module shares the same orchestration layer, the same audit trail, and the same integration fabric — so a win in one surface lands everywhere.

cascade · surface

Module · 01

Six lenses, not one.

Economic, social, technological, political, legal, environmental. Each agent reasons within its domain; the tree composes their outputs into a coherent map.

cascade · runtime

Module · 02

First, second, third order.

Branches fork. A first-order consequence produces second-order effects, which produce third-order ones. Zoom, pan, collapse — the depth is navigable.

cascade · integrations

Module · 03

Every node is reasoned.

Click a node: agent reasoning, source citations, confidence. No black boxes — the tree is auditable all the way down.

cascade · audit

Module · 04

Compare two scenarios side by side.

Run "status quo" and "proposed change" against the same prompt. Diff the trees; see what each decision uniquely causes.

Under the hood

The surfaces the product is made of.

The tree

A D3 canvas you can navigate.

Zoom, pan, drill into any node. Branches fork by order; depth is explicit. Export as JSON for downstream tools or as PNG for a deck.

D3 tree · zoom and pan
Per-node reasoning panel
JSON · PNG export

cascade · the tree

Streaming agents

Watch the agents think.

Six SSE-streamed agent cards show progress in real time — who's generating which order of consequence, what they've cited, where they're converging.

SSE streams from six agents
Live progress per domain
Inter-agent citations highlighted

cascade · streaming agents

Compare mode

Two futures, one diff.

Run two scenarios; Cascade overlays the trees. Unique branches tint; shared branches deduplicate. The decision reveals itself.

Paired scenario runs
Unique vs shared branch overlay
Export paired diff as JSON

cascade · compare mode

Specifications

The shape of the product.

What comes in, what goes out, what it runs on. Nothing hidden in a sales deck.

Agents: Economic · Social · Technological · Political · Legal · Environmental
Depth: First-, second-, third-order consequences
Streaming: Server-Sent Events · live per-agent progress
Output: Interactive D3 tree · reasoning per node · citations · confidence
Stack: Next.js 16 · React 19 · Tailwind 4 · Python 3.12 backend

Pairs well with