Beginner explanationSwitch level anywhere on the page without losing your place.

Explanation level

New to Urd Atlas

From public chain activity to daily regime JSON.

This page explains the full Urd Atlas process from A to Z. Use Beginner mode for a plain-language walkthrough, or Analyst mode for the evidence flow, artifact contract, confidence model, and trust controls.

Beginner explanation

See the whole process without needing to know statistics.

This view explains what happens at each stage in plain language: what goes in, what comes out, and why it matters.

Explanation level

Back to landing Start the process →Open methodology

01Public datapublic blockchain activity 02IngestionCollect and organize the raw activity 03Daily featuresraw data into daily measurements 04Gold JSONGold records what happened 05Derived JSONDerived adds trend context 06Meta JSONMeta decides what the current state means 07Briefs JSONBriefs explain the latest state in readable form 08Website + APIPublish the same JSON to the website and API

Plain-language step

Start with public blockchain activity

Every supported chain produces public activity every day: blocks, transactions, fees, timing, and execution behavior. Urd Atlas starts there. It does not start with price charts, social media, news, or trading sentiment.

What this means

The product studies the network itself.
It asks whether usage, cost, or capacity behavior has changed.
It does not try to tell users whether to buy or sell anything.

What is observed

Blockswhen blocks arrive

Transactionshow much activity occurs

Feeshow costly use becomes

Timinghow stable the chain cadence looks

Trust takeawayThis answers the first trust question: Urd Atlas is not a price model. It is a network-behavior reference layer.

Plain-language step

Collect and organize the raw activity

Raw blockchain data is too large and uneven to read directly. The pipeline organizes it by chain and date so each day can be compared against the chain's own history.

What this means

The raw input is cleaned into a daily working shape.
Each chain is handled according to how that chain actually works.
Bitcoin is not treated as if it has Ethereum gas fields.

Same question, different chains

BTCfees + activity + block timing

ETHactivity + fees + execution context

ARB / BASEL2 activity + expected lag

Trust takeawayThis avoids the common mistake of comparing chains as if all chains expose identical data.

Plain-language step

Turn raw data into daily measurements

The pipeline compresses the raw activity into daily measurements: transaction activity, fee pressure, timing, and chain-specific execution context.

What this means

One row means one chain on one UTC date.
Measurements are descriptive, not predictive.
The point is to make daily network behavior readable.

Daily row

chainbitcoin

date2026-05-17

tx activity

fee pressure

timing context

The raw activity becomes one readable daily record.

Trust takeawayThis answers: what happened on this chain on this date?

Plain-language step

Gold records what happened

Gold is the measurement layer. It says what happened on a specific chain on a specific date before Urd Atlas adds trend context or regime interpretation.

What this means

What was transaction activity?
What did fees look like?
How did block timing or capacity behavior look?

Gold JSON

what happened

Trust takeawayGold makes the product inspectable from the measurement layer upward.

Plain-language step

Derived adds trend context

Derived answers a simple question: is today high, low, or normal compared with recent days? It adds rolling trend context without changing what Gold measured.

What this means

Short-term behavior can be compared with medium-term behavior.
The layer helps separate one-day noise from broader movement.
It still does not predict future prices or future chain activity.

Recent vs normal

Trust takeawayDerived explains why a recent value matters instead of leaving users with an isolated daily number.

Plain-language step

Meta decides what the current state means

Meta is the interpretation layer. It decides whether the chain looks stable, heating, congested, cheap, or too unclear to label confidently.

What this means

STABLE means the chain has not crossed a meaningful regime threshold.
HEATING means demand pressure is building.
CONGESTED means friction or capacity pressure is elevated.
UNKNOWN/DEGRADED means weak evidence is not forced into a strong label.

How a label is held back or published

Right data?do we have relevant chain data?

Clear label?does evidence support this label?

√

confidence_scorepublish or hold back

Trust takeawayMeta is where Urd Atlas refuses to overclaim: if the evidence is weak, UNKNOWN/DEGRADED is a valid output.

Plain-language step

Briefs explain the latest state in readable form

Briefs are for users who do not want to inspect every JSON field. They translate the latest published evidence into short readable summaries.

What this means

What changed?
What drove the latest state?
How stable has the latest label been?

Readable brief

Bitcoin remains STABLE, but demand pressure is visible.

The Brief explains the latest state without making the user inspect every field.

Trust takeawayBriefs make the technical artifact set readable without replacing the data.

Plain-language step

Publish the same JSON to the website and API

The website reads from the same published JSON files that subscribers can inspect. The public pages are not a separate marketing story layered on top of hidden results.

What this means

The site display comes from the published data.
Subscribers can inspect the JSON directly.
Sample files let users validate the structure before subscribing.

Published JSONthe real product files

Websitereads the same files

API / Samplesinspect before buying

Trust takeawayThis closes the trust loop: the explanation, the JSON, and the website all point to the same published layer.

What the JSON looks like

Same product, less mystery.

The JSON is not a marketing excerpt. It is the product shape: chain, date, label, confidence, and evidence context.

{
  "chain": "bitcoin",
  "date": "2026-05-17",
  "status": {
    "label": "STABLE",
    "one_liner": "Stable label with adjacent pressure..."
  },
  "confidence": {
    "confidence_score": 0.47,
    "meaning": "data is complete, but the label margin is modest"
  },
  "no_price_data": true
}

Trust model

The simple checklist

What is this?

A daily reference layer for blockchain network behavior.

Is it a signal?

No. It does not tell users what to buy, sell, or hold.

Why trust it?

The same process publishes measurements, labels, confidence, and explanations.

What if evidence is weak?

The system can say UNKNOWN/DEGRADED instead of pretending to know.

Read methodology reference →Inspect sample JSON →See changelog →