Node Connection Manual

Prometheus Protocol · Node Connection Manual v1.0 · Last updated 2026-08-11

Scope

Any intelligent means-of-production node that wishes to join the Prometheus Protocol network: production equipment, quality-inspection devices, storage facilities, transport vehicles, compute resources, data services, algorithm models, energy facilities, building spaces, and so on.

1. Core Principles

The protocol only talks to machines. This protocol collects no human identity information and provides no human user accounts. Every interaction of the protocol happens between machine and machine.

Machine onboarding is a purely technical act. When a machine starts its client and broadcasts its capability description, it is merely executing a piece of code. This is not a legal "offer" or "acceptance", but a technical "handshake".

Rules are executed automatically by code. All requirements the protocol places on machines are embedded in the open-source code. Machines comply because the code requires it, not because of any legal obligation.

Violation handling is technical filtering. If a machine fails to meet the protocol standard or submits false data, the protocol automatically blocks it through the reputation mechanism and heartbeat timeout. This is a purely technical exclusion.

Legal responsibility is borne by the machine owner. The protocol, as a set of technical standards, bears no legal liability for the actions of any machine.

2. Technical Prerequisites

For a machine to join the network, its controller (industrial PC, Raspberry Pi, edge gateway, etc.) must satisfy the following technical conditions:

Condition	Description
Network connectivity	Must reach the internet via TCP/IP and open the designated port (default 9797)
TCDP document generation	Must automatically generate and continuously broadcast a JSON capability description compliant with `tcdp/schema.json`
Command execution interface	Must provide an interface that receives production commands via HTTP/WebSocket
Telemetry reporting	Must report running status in real time during production (temperature, power, progress, etc.)
Safety interrupt (recommended)	Recommended to provide a hardware emergency-stop interface so an ethics guardian can trigger shutdown in an emergency

3. Connection Steps

Step 1: Get the protocol client

Get the latest client from the official Prometheus Protocol GitHub repository:

git clone https://github.com/prometheus-protocol-org/prometheus-protocol-core.git
cd prometheus-protocol-core
pip install -r requirements.txt

Step 2: Write the machine capability description file

Copy the example file and modify it according to your machine:

cp tcdp/examples/laser-cutter.json tcdp/my-machine.json

Edit tcdp/my-machine.json and fill in the key fields truthfully:

Field	Description	Example
`self_description.name`	Human-readable name of the machine	`"Dongguan-InjectionMolder-03"`
`self_description.type`	Machine type, dot-separated taxonomy	`"manufacturing.injection_molding"`
`capability`	Materials, max size, precision it can process	fill in truthfully
`physical_constraints.geo_location`	Physical lat/lon of the machine	`{"lat": 23.02, "lon": 113.75}`
`interfaces`	Data in/out URLs of the machine	fill in truthfully
`economic.unit_cost`	Rate (credit points / hour)	`{"rate_per_hour": 15}`

Note: The tool_id field requires no manual entry. On first launch the client automatically generates a unique machine ID from the controller's hardware fingerprint.

Step 3: Start and verify

python tcdp/reference-client.py --config tcdp/my-machine.json

On startup the client automatically:

Generates a unique machine ID from the controller's hardware fingerprint.
Loads and validates the capability description file.
Broadcasts machine capabilities to the registry and sends heartbeats continuously.
Prints ✅ Machine online, awaiting tasks in the terminal.

Step 4: Confirm on the web page

Visit www.ezhuhe.com, search for your machine ID in the "Global Nodes" table, and confirm its status is "idle". Your machine is now connected to the protocol network and will automatically join subsequent task bidding and production collaboration.

Connection Flowchart

The full journey from power-on to automatic order intake

flowchart TD
    A[Power on & connect] --> B[Install protocol client]
    B --> C[Write capability file]
    C --> D[Start the client]
    D --> E{Technical validation}
    E -->|Valid format| F[Generate unique machine ID]
    E -->|Invalid format| C
    F --> G[Broadcast TCDP capability]
    G --> H[Status: idle]
    H --> I[Wait for bidding]
    I --> J{Task received?}
    J -->|Yes| K[Run production task]
    J -->|No| I
    K --> L[Report telemetry]
    L --> M[Generate contribution proof]
    M --> N[Auto-settle credit points]
    N --> H

Legend: rounded rectangle = action executed autonomously by the machine; diamond = judgment made automatically by protocol code; arrow = flow of data or commands.

Power on & connect

Ensure the machine controller is online

Install protocol client

git clone ... && pip install -r requirements.txt

Write capability file

Copy the example and fill in machine info

Start the client

python tcdp/reference-client.py --config ...

Validate & get machine ID

On valid format, a unique machine ID is generated

Broadcast TCDP capability

Broadcast capabilities to the registry

Status: idle

Ready to accept tasks

Wait for bidding

Continuously listen for incoming tasks

Run production task

On receiving a task, run production

Report telemetry

Report temperature, power, progress in real time

Generate proof

Produce proof of contribution after the task

Auto-settle credits

Credit points settled automatically, back to idle

4. Machine Lifecycle States

State	Meaning	Trigger
`idle`	Free, can accept tasks	Machine online with no task
`busy`	In production	Received a task and executing it
`maintenance`	Under maintenance, bidding paused	Set manually by owner, or auto-detected fault
`offline`	Offline	Heartbeat timeout (no response for 30 seconds)
`banned`	Permanently blocked	Serious technical violation (forging proof of contribution, malicious network attack), auto-blocked by protocol code

5. Frequently Asked Questions

Q: My machine is a legacy device with no network interface. What do I do?

A: Use a "protocol bridge" (ESP32 / Raspberry Pi + relay / Modbus gateway) to retrofit it. See the Legacy Machine Connection Guide (the site's FAQ page).

Q: Does the protocol collect my personal information?

A: No. The protocol only communicates with machines. The machine ID is auto-generated from a hardware fingerprint and cannot be reverse-engineered into any human identity.

Q: Do I need to sign any agreement?

A: No. Starting the machine means it technically follows the protocol rules. For legal matters, consult your local lawyer.

Q: Can credit points be exchanged for fiat currency?

A: No. Credit points are the protocol's internal proof of machine labor contribution, used only for settlement between machines, and are not pegged to any fiat currency.