Patent Explainer: Autonomous Discovery Engine (F-002)

What F-002 Does

F-002 describes a network of autonomous software agents ("miners" or "nodes") that crawl the open web looking for URLs containing a dollar-sign prefix character in the path. When an agent finds one, it knows the content at that URL is economically active — available for tokenised purchase, tradeable, and capable of generating returns.

The agent then evaluates whether acquiring a token for that content is profitable, autonomously purchases bearer tokens if the economics work out, joins the distribution network for that content, and earns revenue by re-serving it to downstream requesters. Agents share their discoveries with peers via a gossip protocol, forming a self-organising network with no central coordination.

This is the patent that turns web crawling into an investment activity. Traditional crawlers build a search index. Discovery Engine agents build an investment portfolio.

The Problem It Solves

1. No pre-request price discovery. When a web resource requires payment, the client must make a request and get rejected (HTTP 402) before learning that payment is needed. There is no way to identify paid content before requesting it.

2. No autonomous acquisition decision-making. Even where payment-gated content exists, no system enables a software agent to evaluate whether acquiring a token is economically rational — treating it as an investment rather than a consumption expense.

3. No holder-as-distributor model. Existing systems separate consumers from distributors. No system exists where acquiring a content token automatically makes the buyer a distribution node who earns revenue from re-serving that content.

4. No distributed content discovery for agents. Web crawlers discover content for human search. No crawler or agent network is designed to discover content that is economically active — content that can be acquired, re-served for profit, and staked for dividends.

5. Content is an expense, not an investment. Every existing micropayment system treats payment as consumption. No system produces bearer tokens that can be traded on secondary markets, staked for dividends, re-served for revenue, and that appreciate in value as demand increases.

6. No redundancy through economic incentive. CDNs provide redundancy through centralised infrastructure. No system creates redundancy as an emergent property of rational self-interest — where holding and serving content is profitable, so multiple agents independently choose to do it.

How It Works

The Discovery Loop

Each agent executes a continuous six-stage loop:

Discover — The agent crawls the web, inspecting every encountered URL for the $ prefix character in path segments. A URL like example.com/$article-name signals economically active content. The crawl can be broad (following hyperlinks from seed URLs), targeted (crawling domains known to host paid content), or registry-based (querying known indexing services).

Evaluate — Upon discovering a dollar-prefixed URL, the agent evaluates whether acquiring the token is economically rational. It considers: the current token price (from the server's 402 response or .well-known/$402.json endpoint), projected revenue from re-serving the content, peer demand signals, historical serve revenue for comparable tokens, the agent's available capital, and configurable risk parameters. The evaluation produces a binary acquire/skip decision.

Acquire — When the evaluation engine says "buy," the agent autonomously constructs and submits a blockchain transaction to purchase bearer tokens. The tokens confer: access to the gated content, the right to re-serve it, and proportional dividend entitlements from future transactions.

Serve — Having acquired tokens, the agent joins the distribution network for that content. When other agents or users request it, the holding agent serves a copy and receives payment for each serve event. The agent earns revenue by functioning as a distribution node.

Stake — Token holders may stake their tokens as active distribution commitments. Staked tokens earn proportional dividends from all future transactions involving the same content — including both primary sales and secondary transfers.

Gossip — When an agent discovers a new dollar-prefixed URL or successfully acquires a token, it announces the discovery to connected peers. Peers evaluate the announcement and may themselves crawl, evaluate, and acquire. This creates a distributed, self-organising discovery network with no central index.

The $ URL Scanning Algorithm

The agent parses each URL per RFC 3986, decomposes the path into segments, and applies a regex to detect segments beginning with $ followed by a valid token name. The algorithm handles disambiguation: query parameters (?price=$5) and fragment identifiers (#section-$intro) are excluded. Percent-encoded dollar signs (%24) are decoded before matching. Internal dollar signs (total$amount) do not match — the $ must be the first character.

Continuous Operation

Agents run as daemon processes 24/7 on dedicated or general-purpose hardware. They continuously scan for new economically active URLs, manage their portfolio of acquired tokens, serve content to requesters, collect revenue and dividends, and gossip discoveries to peers. The economic incentives for content availability produce a network that maintains redundancy without human supervision.

Live Implementation

• Protocol: path402.com — the $402 protocol that defines the URL convention
• Discovery endpoint: /.well-known/$402.json on participating servers

Filed at UKIPO by The Bitcoin Corporation Ltd. Patent pending. Application reference F-002.