Whitepaper / Start Here

Recognition and Traction

This chapter records the outside recognition Ghast AI has received, with the source for every number in it. It is not a forecast, a customer list, or evidence that Ghast Protocol is live. Protocol is in development. Capability commerce does not exist yet, so nothing here can be read as revenue from it.

Hackathon result

Ghast AI took first place at the 2026 0G APAC Hackathon, out of 293 submissions from 1,145 builders. See the 0G Labs announcement and official hackathon recap.

The hackathon was run by 0G Labs with HackQuest. The organiser’s recap is the source for the submission count and the builder count. Second and third place, and the excellence awards, are named in the project’s own account of that result. This chapter does not add a ranking beyond first place out of 293 submissions.

The entry that won and the desktop application a person installs today are not the same feature set. The winning entry was described by the organiser as a browser-native assistant using the 0G stack. Ghast Workbench today is the desktop workspace for Mac and Windows, with Chat, Coding, and Agent Team. The site’s hackathon article keeps the sources, the date, and that distinction together. Use it when you need the announcement-level write-up. Use this chapter when you need the result inside the whitepaper.

A first-place hackathon result is recognition of a submission. It does not mean a marketplace launched, a token was issued, or a partnership was formed beyond what the organiser published. This whitepaper does not add partners, customers, or contract values that are not already stated.

What the judges looked at

A hackathon judges what a team built in a fixed window and shows on a fixed date. The judges saw a working submission, its demonstration, and the team’s account of what it did. They did not audit a business, a user base, or a production deployment, and the result does not claim they did.

That is why the result is recorded as a date and a count, not as a rating of the company. “First of 293” is a fact about one competition in 2026. It says the submission was judged strongest in that field. It does not say how the product performs today, which models it supports, or what it costs to run. Those questions are answered by the product itself and by the user guide, not by a ranking.

The result is also not a delivery record in the sense this whitepaper uses. Reputation and Delivery History describes trust that would belong to one Capability and be built from purchases, evidence and outcomes. A competition placing has none of those parts. It cannot be attached to a Capability, and it does not raise the standing of any service a Creator might publish later.

Product usage figures

This whitepaper does not publish Ghast Workbench usage figures. Accounts, paying users, active users, inference volume and compute consumed are operating data. They are not part of the design this whitepaper describes, and they would not prove any part of it.

The evidence that Workbench is live is Workbench. It can be downloaded and installed for Mac and Windows, and the user guide documents Chat, Coding, Team, models and the settings a person can check. Ghast Workbench describes the product in this whitepaper, and Infrastructure Overview places it as the live layer beneath a Protocol that is still in development.

If a usage number about Ghast circulates without a source on this site, treat it as unofficial. This chapter will not adopt a figure from a third party, and it will not round one into a claim.

What the recognition is allowed to support

The result supports a narrow statement. An independent hackathon, run by 0G Labs with HackQuest, judged the project first among 293 submissions in 2026. That is consistent with Workbench being the live product, and with Ghast AI building on private AI work before any Protocol exists.

It does not support a statement that Ghast Protocol is operating, that Capabilities are being purchased, or that $GHAST functions are live. Capability Economy describes a proposed market. Token Allocation describes how a proposed supply would be divided. Neither chapter inherits the hackathon result as proof.

A ranking also does not verify a promise. Verification, in Evidence and Verification, would compare Delivery Evidence to one Defined Claim. Judges comparing submissions is a different act. A first place says nothing about whether a future Capability met the claim a buyer paid for.

The result does not identify customers. This chapter names no company as a client and no person as a buyer of a service. Treating hackathon recognition as a portfolio of design partners would invent relationships the organiser did not state.

How identity and history treat this result

Identity and Provenance separates the identity of a Capability from the reputation of the people or company behind it. The hackathon result is a fact about the project. It is not a Capability identity, and it does not carry over into one.

Security and Risk sets out the risks a buyer of a future Capability would carry. Recognition does not reduce any of them. A buyer of a future Capability would rely on its claim, its evidence and its delivery history, not on how the team placed in a competition.

How to read this chapter next to the product

Install Ghast Workbench, and follow the user guide for Chat, Coding, Team, and models, when the goal is to use the live product. Read System Overview and then the protocol chapters when the goal is the design that would sit on top of private work. The news account of the 2026 result gathers the hackathon sources in one article. Keep the date attached. A June hackathon result answers one question. It is not a Protocol launch.