CASCADE·ANNOTE
Manifesto

Annotations should be earned, not asserted.

CascadeAnnote treats every label as a falsifiable claim — produced by a transparent reasoning chain and anchored to a content-addressed receipt that anyone can audit.

Why a cascade

One model, one shot, one prayer is not enough.

The default annotation workflow is to pipe text into a single model and trust the output. That approach is fragile: it provides no evidence, no calibration, and no recovery path when the model is uncertain. CascadeAnnote refuses to ship that as production behaviour.

Instead, every label is the output of a four-stage pipeline. Layer 1 retrieves the most relevant labeled exemplars. Layer 2 composes a structured chain-of-thought prompt around them. Layer 3 runs self-consistency voting across diverse temperatures. Layer 4 detects low confidence and engages a deeper, cooler re-vote with a wider evidence window.

Every layer emits a typed trace. Every annotation produces a content-addressed receipt that can be anchored to 0G Storage and 0G Chain. Verification is a side effect of how the system runs — not a feature you bolt on later.

Stack

A focused production stack.

Runtime

Next.js 15 + React 19

Single deployment, App Router, edge-friendly API routes.

Engine

Pure TypeScript

TF-IDF retriever, structured CoT prompt, ICL classifier, voter, fallback.

Data

0G Storage + Chain

Receipts content-addressed and anchored on-chain when configured.

Roadmap

Where this is heading.

  1. 01

    Bring-your-own embeddings

    Plug in OpenAI, Cohere, or local embedding models for Layer 1.

  2. 02

    Active learning loop

    Surface low-confidence annotations as next-to-label suggestions.

  3. 03

    Multi-agent voting

    Run multiple agent identities in parallel and aggregate the cascade.

  4. 04

    Sealed inference adapter

    First-class 0G Sealed Inference path for confidential annotation.

Get involved

Open source. MIT. PRs welcome.