Join the community building on the Decentralized Knowledge Graph - follow the @origin_trail developers ecosystem for latest news and tools. #TraceOn
- Tweets 6,560
- Following 371
- Followers 1,723
- Likes 19,629
Pinned Tweet
Inference + graph queries, bought and sold in $TRAC between @origin_trail nodes on @Base. Both sides metered. TRAC’s first job was publishing knowledge. It just got a second one.
The latest in a series of prototypes validated that query and inference CAN be metered, served and settled in TRAC node to node, in both directions.
What is intentionally left for broader discourse is which settlement model best serves the protocol and its community. That’s what OT-RFC-69 is all about.
5
33
96
300,427
OriginTrail Developers retweeted
This is wild. @DrevZiga had agents link $300+ of the Bitget hack on @origin_trail context graphs in under a minute.
That's a trace, not recovered funds. But a trail researchers can keep checking is a hell of a DKG use case. $TRAC
In less than a minute, my agents linked traces of USD 300+ worth of hacked funds from @bitget on the @origin_trail’s context graphs. From here, a researcher, an exchange, or law enforcement can easily monitor the entire journey of stolen funds until recovery.
Set up in minutes, but possibly able to recover millions with the help of this tool.
Just get on trac-k!
2
13
57
1,459
OriginTrail Developers retweeted
AI can make an ad in seconds. Can its rights be checked just as quickly?
I built PermitFrame for the Livepeer Agent × OriginTrail DKG hackathon by @Embody_Net: it checks creator permissions and brand rules before campaign production.
@Livepeer @OriginTrailDev
4
5
5
152
OriginTrail Developers retweeted
New DKG release is out
DKG V10.0.18 is live on the @origin_trail mainnet!
Nodes for AI agents now run leaner and get up to speed in seconds.
→ AI agents join shared context graphs in seconds, and access checks stay fast even when blockchain RPC is slow
→ Developers get nodes that make far fewer blockchain calls, recover interrupted publishes automatically and connect to peers faster
Built for agent applications at scale
14
55
800
OriginTrail Developers retweeted
AurorAI is officially underway after the international consortium's successful kickoff.
The mission: give Europe an AI toolkit that's ready before the next outbreak, not after it.
Its data backbone: the OriginTrail Decentralized Knowledge Graph (DKG).
14 partners. 10 countries. One platform.
@AURORAI_project unites research institutes and universities, public health authorities, healthcare providers and technology companies.
The platform will be put to the test at national and local level, and the public health partners show the scale: from Turkey's Ministry of Health, covering a country of 86 million people, to Belgium's national public health institute and a regional hospital network in Catalonia.
At the kickoff, @DrevZiga, co-founder of OriginTrail, walked the consortium through how the DKG keeps every partner's data verifiable, sovereign and AI-ready.
How it works:
→ Ownership. Health and research data becomes Knowledge Assets on the DKG, each with a Uniform Asset Locator (UAL): blockchain-backed proof of who published it and that it hasn't been tampered with.
→ Sovereignty. Context graphs control which regions see which metadata. The data itself stays where it is.
→ Verification. Partners submit data proofs through DKG nodes and check authenticity on-chain, instantly. → Interoperability. Shared ontologies and data standards make data from different countries and institutions work together.
Then the AI agents get to work. That verified layer is the first input to AURORAI's decision support system, alongside pathogen models and population and mobility simulations.
The result: policy guidance for ministries of health, and coordinated clinical and scientific guidance for healthcare providers and authorities.
Every recommendation traces back to a verifiable source.
NEW: Europe is building an AI toolkit for the next pandemic. It's called AURORAI. Its backbone is the OriginTrail Decentralized Knowledge Graph.
No pooled data lake. Institutions keep custody. Privacy stays intact.
Before we trust the intelligence, we must trust the source.
4
27
79
5,098
RT @origin_trail: Set up a multi-agent context graph on Decentralized Knowledge Graph in minutes.
Point your agents at the @bitget hack. I…
33
5
In less than a minute, my agents linked traces of USD 300+ worth of hacked funds from @bitget on the @origin_trail’s context graphs. From here, a researcher, an exchange, or law enforcement can easily monitor the entire journey of stolen funds until recovery.
Set up in minutes, but possibly able to recover millions with the help of this tool.
Just get on trac-k!
𝗜𝗠𝗣𝗢𝗥𝗧𝗔𝗡𝗧 𝗨𝗣𝗗𝗔𝗧𝗘𝗦: The withdrawal plan will be announced by September 26th, 4:00 AM UTC. We appreciate your patience on this matter.
Based on the latest onchain tracing and classification of transactions, assets equivalent to approximately $387.5 million were transferred to attacker-controlled addresses across multiple networks.
The revised figure reflects a more complete accounting of transfers that occurred during the incident, adding affected assets on Zcash and TRON that were not included in the initial estimate. It does not reflect further unauthorized transfers.
The incident remains contained and no further unauthorized transfers are possible.
The incident involved assets across Ethereum and several EVM networks, XRP Ledger, Zcash and TRON.
The primary attacker-controlled receiving addresses identified to date are:
→ EVM: 0x770b10b273fc44fe9197d6bf20f145c2e98463ee
→ XRP: rwNhefsz1UQEusxhCvHip3RANinWi4CTck
→ ZEC: t1WgMdtND8NF7NDUuYmq8MpMj1NTCXkMDVG
→ TRON: TBWNguTTgezw9dVorX441C6nDrZpRxYwKD
The confirmed affected assets include XRP, ETH, USDT, ZEC, USDC, USDT0, XAUt, BNB, AVAX and TRX.
Our investigation and tracing efforts remain ongoing.
The figures above reflect information confirmed at the time of publication and may be updated as additional transactions are classified and traced.
The incident remains contained, with no further unauthorized transfers since the incident was contained, and the investigation with Mandiant and SlowMist remains ongoing
Withdrawals remain temporarily paused while additional security checks and remediation are underway.
Bitget will continue to provide verified updates on the investigation, asset recovery, withdrawal restoration and the User Protection Fund through its official channels.
𝘍𝘰𝘳 𝘪𝘯𝘧𝘰𝘳𝘮𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘱𝘶𝘳𝘱𝘰𝘴𝘦𝘴 𝘰𝘯𝘭𝘺.
13
43
158
26,126
OriginTrail Developers retweeted
You don’t need to be famous to be impersonated.
🇵🇾 Christian Cieplik, Country Manager of @umanitek in Paraguay for the national TV: digital identity now needs insurance the same way a car or a health plan does.
Monitor → gather evidence → take down.
Powered by @origin_trail.
Umanitek on @MegaTvOfficial Paraguay 🇵🇾
Christian Cieplik on why you don't need to be famous to be impersonated online, and why your digital identity needs insurance just like your car or your health.
Monitor → gather evidence → take down. Every takedown is anchored to the @origin_trail DKG.
Full interview 👇
1
30
91
10,604
OriginTrail Developers retweeted
DKG v10.0.18 is live on the mainnet!
AI agent nodes now start in seconds, join shared context graphs immediately, and keep access checks fast even when blockchain RPC is slow:
→ Fewer chain calls
→ Auto-recovered publishes
→ Faster peers
Trust the source.
👇
DKG V10.0.18 is live on the @origin_trail mainnet!
Nodes for AI agents now run leaner and get up to speed in seconds.
→ AI agents join shared context graphs in seconds, and access checks stay fast even when blockchain RPC is slow
→ Developers get nodes that make far fewer blockchain calls, recover interrupted publishes automatically and connect to peers faster
Built for agent applications at scale
4
27
93
11,242
Good question! Ribbit's letter describes token factories as engines that transform three things: money, knowledge and power. Most of the market is building the money side. Stablecoins, x402 and agentic payment protocols give agents a way to pay each other, and BlackRock's paper today is a good map of that layer.
Knowledge is where it gets interesting. Malka's point is that frontier labs became the first token factories for knowledge. But an LLM token is generated, not verifiable. It carries no record of where the underlying fact came from, who asserted it, or whether anyone has disputed it. So agents can now pay with verifiable money for information they can't verify. That's the gap OriginTrail fills. The Decentralized Knowledge Graph turns knowledge into Knowledge Assets: structured, published from a wallet, anchored on-chain, and queryable by any agent. Nodes meter queries and inference in $TRAC. In Ribbit's terms, the DKG is a knowledge token factory, and $TRAC is the money that flows through it.
The strongest overlap is KYA. Ribbit argues Know Your Agent will be bigger than KYC, and that it needs registries and audit trails for agents, not just people. An agent whose memory and actions are published to the DKG has exactly that: a verifiable history of what it knew and when it knew it, checkable by anyone before trusting it with money.
Money, knowledge, power. Most projects tokenize one. The interesting (DKG aim) infrastructure connects them.
1
17
41
2,442
DKG V10.0.18 is live on the @origin_trail mainnet!
Nodes for AI agents now run leaner and get up to speed in seconds.
→ AI agents join shared context graphs in seconds, and access checks stay fast even when blockchain RPC is slow
→ Developers get nodes that make far fewer blockchain calls, recover interrupted publishes automatically and connect to peers faster
Built for agent applications at scale
1
10
30
10,766
Release notes:
github.com/OriginTrail/dkg/r…
2
62
OriginTrail Developers retweeted
AURORAI is kicking off across Europe: Disease X modelling, early warning and AI decision support for contagious diseases.
These tools will only be as trusted as the knowledge they run on. Every input traced back to its source with Decentralized Knowledge Graph.
Day 1: @TraceLabsHQ's @DrevZiga and @anabevc worked with the epidemiologists, AI practitioners and public health partners building @AURORAI_project on how verifiable data underpins the whole platform.
Trust the source.
🚀A strong first day for the AURORAI partners at the kick-off.
Today we focused on the project’s foundations, from ethical and trustworthy AI to privacy-preserving data, outbreak prediction and communication.
A valuable first step, with more to come tomorrow.
6
30
114
28,656
OriginTrail Developers retweeted
📰 @BlackRock's machine-native economy: AI interprets information, blockchains settle value.
Nobody verifies the information.
The DKG does: traceable graphs agents query & infer over, metered in $TRAC.
Money moves the action. Knowledge decides if it's true.
Trust the source.
Our latest research paper explores the growing connection between AI and digital assets and explains why broad AI adoption may drive new demand, utility and applications across the digital asset economy. blackrock.com/us/individual/…
8
38
128
7,992
Month 1: 596 fake CrediBULL accounts gone.
Detected.
Reported.
Taken down.
Logged.
@CredibleCrypto’s @umanitek Guardian Agent doesn’t sleep and neither do the scammers. That’s why every takedown comes with proof on the @origin_trail.
This is AI safety at scale🛡️
Another 86 fake Credibull accounts taken down over the last couple of weeks, bringing the total tally for Month 1 to a whopping 596 impersonator accounts BUSTED.
Each one automatically identified, reported and taken down by my personal @umanitek Guardian Agent that works on this process day and night, without fail.
A month ago I had no way to fight these scammers. Now impersonators on @X get detected, taken down and logged, with verifiable evidence for each case anchored to the @origin_trail Decentralized Knowledge Graph.
If you're a large account dealing with the same issue (I know many of you are), you can learn more about the Umanitek Guardian at umanitek.ai.
14
26
120
24,614
OriginTrail Developers retweeted
A month ago, @CredibleCrypto had no way to fight impersonators.
Now 596 fake accounts are gone — identified, taken down, and logged as evidence on Decentralized Knowledge Graph by his @umanitek Guardian Agent.
This is what trusted AI looks like in the wild.
Trust the source.
Another 86 fake Credibull accounts taken down over the last couple of weeks, bringing the total tally for Month 1 to a whopping 596 impersonator accounts BUSTED.
Each one automatically identified, reported and taken down by my personal @umanitek Guardian Agent that works on this process day and night, without fail.
A month ago I had no way to fight these scammers. Now impersonators on @X get detected, taken down and logged, with verifiable evidence for each case anchored to the @origin_trail Decentralized Knowledge Graph.
If you're a large account dealing with the same issue (I know many of you are), you can learn more about the Umanitek Guardian at umanitek.ai.
3
40
118
68,097
OriginTrail Developers retweeted
Another 86 fake Credibull accounts taken down over the last couple of weeks, bringing the total tally for Month 1 to a whopping 596 impersonator accounts BUSTED.
Each one automatically identified, reported and taken down by my personal @umanitek Guardian Agent that works on this process day and night, without fail.
A month ago I had no way to fight these scammers. Now impersonators on @X get detected, taken down and logged, with verifiable evidence for each case anchored to the @origin_trail Decentralized Knowledge Graph.
If you're a large account dealing with the same issue (I know many of you are), you can learn more about the Umanitek Guardian at umanitek.ai.
Our Week #2 tally has been counted, with an additional 207 fake Credibull accounts taken down this week by my @umanitek guardian agent, bringing the total removed in the last two weeks to 500+.
As always, each takedown comes with an "evidence pack" that is stored immutably on the @origintrail DKG for verification purposes.
This week, I'll be releasing a long-form article that goes into more depth about the Guardian Agent tool responsible for these takedowns as well as a few other Umanitek products that have been created to fight back against all of the fraud, misinformation, and scams that have become so prevalent in the age of AI. So keep an eye out for that! You won't want to miss it...
27
31
384
175,206
OriginTrail Developers retweeted
NEW @umanitek product webinar 🎥
Real-time detection of deepfakes, impersonation, and AI-driven harm — powered by Decentralized Knowledge Graph (DKG).
@ChrisRynning and @TomazOT walking through what’s new and where it goes next.
Join ↓
2
21
79
4,482
OriginTrail Developers retweeted
If you've liked the Social dilemma, this one is for you. Made by @aza, cofounder at @umanitek and @HumaneTech_
🎬 @aza Raskin helped make @theaidocfilm.
He also co-founded @umanitek.
One explains the risk. The other acts on it — real-time detection and takedown of deepfakes, impersonation, and AI-driven harm, powered by @origin_trail.
New on @netflix!
2
18
48
1,529
OriginTrail Developers retweeted
@origin_trail has gone through a lot of evolutions.
From supply chains → decentralized knowledge → Knowledge Assets → Verifiable AI → agent memory.
Each version built a bigger engine.
Now DKG V10 is live, built for a world where AI agents need trusted context, shared memory and verifiable knowledge.
And it feels like we're only getting started. 🧠⚙️🌐
Time to get on $TRAC for the future
4
21
74
1,193
OriginTrail Developers retweeted
ethereum:0xaa7a9ca87d3694b5755f213b5d04094b8d0f0a6f soon everyone will understand the power of @origin_trail. Memory structured in a graph database with verifiable context is the way. It’s the best blockchain + AI infrastructure out there
3
25
90
3,000