ZERO is the operating intelligence layer between humans and capital.
The first wedge is onchain perpetual markets. The broader product turns fragmented capital state, manual monitoring, opaque agents, weak intervention, and stale proof into an inspectable operating layer: intent, agent network, capital state, governance, and proof.
This public repository is not the whole hosted product. It is the open Runtime, Protocol, and Proof substrate that engineers can run, inspect, verify, and extend without custody or a hosted ZERO account.
Operating intelligence for self-custodial capital operations.
This category is broader than execution software and more concrete than a generic AI agent. It says what ZERO coordinates, who keeps control, and why the product can expand from Hyperliquid-first execution into treasury, hedging, multi-venue routing, and verified intelligence.
ZERO makes capital operations inspectable, interruptible, and self-custodial.
Crypto markets are machine-speed and always open. Human attention is not. Most operator failure is not caused by missing another signal; it comes from inconsistent execution, unmanaged risk drift, unclear state, and manual intervention at the wrong time.
ZERO starts from a different premise: capital needs operating intelligence, not another alert feed. The runtime evaluates opportunities, rejects weak actions, records decisions, and enforces safety policy. The operator surfaces show state, replay, and supervised intervention. The proof layer exposes public-safe evidence. The intelligence layer turns verified behavior into a data product.
The operator keeps custody. ZERO provides the discipline layer.
Attention-driven onchain infrastructure:
- Interfaces that reward activity over outcomes.
- Bots that automate entry but underinvest in safety and post-entry control.
- Copy-trading feeds that cannot be audited or adapted to the operator's policy.
- Dashboards that show state without making the system safer.
- Self-custody first.
- Paper mode before live mode.
- Inspectable behavior over vague claims.
- Risk-reducing actions fast, risk-increasing actions deliberate.
- Public proof without leaking secrets.
- Commercial value based on speed, scale, history, reliability, and verified intelligence.
This repository is the open ZERO Runtime, ZERO Protocol, ZERO Proof, and Developers substrate.
Hosted ZERO Studio, ZERO Control, and ZERO Registry are product surfaces outside this repository; the public repo supplies the software and contracts those surfaces can consume.
- ZERO Runtime: the local runtime for autonomous operations, including CLI/TUI tooling, local memory, genesis, research, paper canaries, and evolution gates.
- ZERO Protocol: MCP-compatible schemas, tool permissions, mandates, replay frames, evidence bundles, audit entries, and local verification fixtures.
- ZERO Proof: redacted proof packets, replay/export fixtures, journal roots, verification commands, and public contract schemas.
- Developers: public education starts at getzero.dev/developers; repo-local docs stay focused on running and verifying the open Runtime/Protocol/Proof substrate.
The ZERO CLI is Runtime tooling, and ZERO Evolution is a local Runtime subsystem. They are not separate hosted public surfaces.
Short GitHub description:
Open ZERO Runtime, Protocol, and Proof for self-custodial capital operations.
Extended description:
ZERO is the operating intelligence layer between humans and capital. This public
repo contains the open Runtime, Protocol, and Proof substrate: paper-first
execution, safety gates, local journals, MCP readouts, public proof packets, and
self-custodial live-operation boundaries.
What to avoid:
- "AI entry bot"
- "Guaranteed yield"
- "Hosted custody"
- "Fully trustless AI trader"
- "Set and forget"
Those phrases either shrink the category, create regulatory and trust risk, or misrepresent the safety model.