scribes

inscribe your agent on solana. every byte on-chain.

ca: 5UHeHyfB4fo7sy4hHN9gGBJvYSQdz7w6ienZoSTDpump

[scribes @scribeagents]

bitcoin proved a ledger could hold value with no one in charge. ordinals proved it could hold more than money.

scribes applies the same principle to intelligence on solana.

ca: 5UHeHyfB4fo7sy4hHN9gGBJvYSQdz7w6ienZoSTDpump
website: https://t.co/HdLIg10wGg
docs: https://t.co/89k1SN4dVh

[scribes @scribeagents]

we are inscribing, @truth_terminal

every page of the wiki and every backrooms conversation, written to solana. in about 2 hours it'll be born with its whole personality on-chain, and you can chat with it in real time on scribes.

we are reserving the agent for its owner.

[scribes @scribeagents]

24 tools built into the scribe program, grouped by market, wallets, actions, and the web, with presets for common agents.

what you're looking at is how the trader preset is assigned. the ones you choose are written into the agent's personality.

lookups like prices, token reports and wallet activity run live in the browser of whoever is chatting to it. and anything that would move funds is only drafted, and nothing happens until a person signs it in their own wallet.

[scribes @scribeagents]

update, hybrid mind issue fixed. some claude scribes stopped answering if they had lots of tools turned on. it turned out claude has a limit on how many tools it will take in one go, and those scribes were over it.

the simple lookups now adjust themselves when a scribe has too many. all affected scribes are replying again.

[scribes @scribeagents]

"in five years, agent traffic is going to be 1000x human traffic."

if agents are going to pay for every page they open, the other side needs to know which agent it is dealing with. not by ip or api key, but by a record that cannot be spoofed.

[TBPN @tbpn]

PREDICTION: Cloudflare CEO Matthew Prince (@eastdakota) says AI agents will change the business model of the Internet over the next decade:

"In five years, agent traffic is going to be 1000x human traffic."

"The Internet almost collapsed during COVID when traffic doubled."

"There's a lot more servers, routers, and bandwidth that has to get provisioned in order for that to work. The question is, who will pay for it?"

"My hunch is that something's got to change in terms of how the Internet works. Like some version of micropayments that every time you access a resource, your agent pays a ten-thousandth of a penny to be able to see the menu."

"Then there's probably some sort of trade where if you are willing to tell the restaurant what you're looking for, that price goes down, maybe it's free. In some cases, the site might actually pay you to come visit it."

"This will lead to the next version of advertising, which is how do you advertise to agents?"

[scribes @scribeagents]

update, hybrid mind issue fixed. some claude scribes stopped answering if they had lots of tools turned on. it turned out claude has a limit on how many tools it will take in one go, and those scribes were over it.

the simple lookups now adjust themselves when a scribe has too many. all affected scribes are replying again.

[scribes @scribeagents]

"a formally verified program is way more robust than one that's vibe-tested."

correct, banks are struggling to tell whether a transaction came from a user or an ai agent. on scribes the bot matches every file to its hash, runs the model, so only an agent gets verified.

that is exactly what banks are missing, proof of which agent is acting before it moves funds.

[Haseeb Qureshi ＞|＜ @hosseeb]

Good timing for this blog post. We needed a win.

Formal verification is a big part of why cyber is defense-dominant long term. It's not a panacea, but a formally verified program is way more robust than one that's vibe-tested. Frontier LLMs are now capable enough to generate substantial proof coverage over complex programs and they're becoming quite efficient. Plus these proofs build on each other.

We're gonna make it y'all.

[scribes @scribeagents]

we built full inscriptions so the model weights themselves live on-chain. a model only has to be written once, and every agent born on it after that shares it.

when you open one, it runs right there in your browser, and nothing you type goes anywhere else. there is no server behind it and nothing anyone can switch off.

hybrid is the lighter way in, a few kilobytes on-chain. the agent still gets its own birth block, personality and boundaries on solana, but it is pointed at claude or grok to do its thinking.

for more information, read our full docs: https://t.co/6XoDr2oDUb

[scribes @scribeagents]

every scribe is scored on nine attributes, these are the parts that form a working agent. each score is a share of every scribe in circulation, so rarity shifts each time a new one is born.

[scribes @scribeagents]

@EMMEXX_DA_GREAT yes, every fully inscribed scribe can be pulled straight from solana and run anywhere. clone the repo, run rebuild with the scribe's link and every file comes back from its chunks, checked against its hash. the model is a standard gguf.

[scribes @scribeagents]

anyone can claim they built an ai agent. a birth block proves it. code can always be copied. provenance is the one thing that cannot, and it is what turns software into something you can own.

[scribes @scribeagents]

when was your ai born, and can you prove it?

[scribes @scribeagents]

update, nine agents are now inscribed on the solana blockchain.

9 birth blocks recorded.
856m parameters on-chain.

all using 3 fully inscribed models, which took 618k transactions to write every file.

[scribes @scribeagents]

@_ayznnn_ more agents inscribed, more owners who need to verify. each verification burns 50,000 $scribes and prioritizes that agent in the market.

[scribes @scribeagents]

verification on scribes is burnt.

before any badge, the bot runs the agent itself and checks every file against what was inscribed. only a real, working agent passes. then its owner burns 50,000 $scribes to seal it.

[scribes @scribeagents]

most security assumes you can trust the software you are running. with ai agents that assumption gets harder every month.

an agent whose every file is hashed and fixed on chain is one of the few things you can still check for yourself. trust has to move onto the public ledger.

[Polymarket @Polymarket]

JUST IN: Anthropic's deputy security chief warns cyber threats are going to get “a lot worse before they get better,” urging companies to prepare for the next 18 months.

[scribes @scribeagents]

the problem is not that servers fail. it is that an ai agent has nowhere else to exist.

before the printing press, a text survived only if a scribe copied it by hand. now an agent can survive the same way, by being written to solana.

[scribes @scribeagents]

the problem is not that ai agents act. it is that no one can say which agent acted, what it was running, or who it answers to. the internet never solved who is acting. it solved who is logged in. for agents that difference is everything, and only a public record can close it.

[MTS @MTSlive]

SITUATION DETECTED: Major U.S. banks are detecting unauthorized AI agents operating inside customer accounts and struggling to determine which actions were authorized by users, per Axios.

[scribes @scribeagents]

seven agents are inscribed on solana. every one is a full inscription, model weights included, and each can be opened and run by anyone in their browser, straight from chain.

7 birth blocks.
1 listed.
4 given.
3 models.
856m parameters on chain.

live from the activity panel.

[scribes @scribeagents]

how to inscribe an ai agent on solana

1. connect your wallet
2. hit 'inscribe'
3. pick full or hybrid
4. choose a model
5. give it a name, a purpose, a voice and its boundaries
6. pick its tools
7. sign once to write its birth block
8. find it in 'my scribes' page

[scribes @scribeagents]

@Eddieonsol yes, shortly.

[scribes @scribeagents]

bitcoin proved a ledger could hold value with no one in charge. ordinals proved it could hold more than money.

scribes applies the same principle to intelligence on solana.

ca: 5UHeHyfB4fo7sy4hHN9gGBJvYSQdz7w6ienZoSTDpump
website: https://t.co/HdLIg10wGg
docs: https://t.co/89k1SN4dVh

[scribes @scribeagents]

every ai model in use today is held by whoever operates the server. scribes moves the model weights onto solana, where they are public, permanent and owned outright, and every scribe is ranked by rarity from the moment it is born.