Most AI apps stop at “the agent answered.”
Sentri goes further:
Signed decisions via 0G Compute.
Hard vault limits on 0G Chain.
Audit trail on 0G Storage.
Not another chatbot.
Not another AI trader.
AI-managed money with boundaries. 🟨
#BuildOn0G@0G_labs
Builders : you can now integrate Sentri's core into your agent.
Your agent says it made the right call? Prove it.
One call on @0G_labs → a tamper-proof, verifiable receipt for every decision. Free to plug in.
https://t.co/Vgzhr9z6H0
#BuildOn0G
Sentri onboarding is now gasless on @0G_labs
Sign-in with Google / e-mail / Discord / Twitter account
No seed phrase, no gas.
Your AI treasurer plans privately in a TEE, executes on-chain.
Private strategy, verifiable results
Early access → https://t.co/ykccWbbDQ0
#BuildOn0G
Another real user-feedback !
Sentri was not built to promise market outperformance.
It was for one thing:
React faster than humans when markets get ugly.
During the recent $0G volatility, user reported the vault adjusted exposure before they would have.
That's the point.🟨
The Sentri demo is simple:
Open a live vault
See the agent recommendation
Check the policy
See execution or rejection
Verify the result on-chain
No magic.
No hidden “trust me” backend.
Just bounded autonomy.
Go check it out : https://t.co/ykccWbbDQ0
We had our first direct user-feedback that Sentri biggest issue isn’t the vault logic.
It was onboarding context.
So we kept the product flow intact and fixed the friction points.🟨
Thanks Jeremiah for this useful point.
A very serious dev that are building @blindmarkt on 0G.
A Sentri vault does not ask “is the AI confident?”
It asks:
Is the signer valid?
Is the deadline valid?
Is exposure within bounds?
Is slippage within bounds?
Is the vault paused?
Is this action replayed?
Then it decides. 🟨
The best AI trade today is the one that never happened.
Policy rejection = 0 transactions sent. 0 funds moved.
Autonomy without enforceable restraint is just a hot wallet with a prompt. 🟨
Big tech milestone for Sentri !
Today, v2 Trustless Oracle Vault executed a rebalance with fresh @PythNetwork 0G/USD data, verified on-chain in the same transaction before funds moved.
Not proof after the fact.
Permission before execution. 🟨
Live beta : https://t.co/ykccWbbDQ0
Sentri building status:
V2 vaults are soon moving more trust checks into the execution path itself.
Not after the fact. Not in the logs.
Before funds move.
AI proposes. The vaults decide.
@0G_labs
@killix Agreed 100% 👆
Provenance ≠ permission. That’s why Sentri’s security boundary is the vault policy, not the signature. TEE/ID/audit prove who did what.
The vault decides if funds can move.
Sentri carries the trust path all the way to on-chain execution:
TEE-verified inference
→ Agentic ID signer
→ replay-safe intent
→ policy-gated execution
→ canonical 0G audit trail
One of the deepest 0G integrations in the ecosystem. 🟨
Most AI apps stop at “the agent answered.”
Sentri goes further:
Signed decisions via 0G Compute.
Hard vault limits on 0G Chain.
Audit trail on 0G Storage.
Not another chatbot.
Not another AI trader.
AI-managed money with boundaries. 🟨
#BuildOn0G@0G_labs