Nava AI (@navaai) header
Nava AI (@navaai) profile picture

Guardrails for the agentic economy || Join our waitlist for private testnet

Joined January 2025
Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)Photo by Nava AI (@navaai)
AI agents are starting to move real money. The problem: there’s still no independent check between what an agent intends to do and what actually gets executed. Today, that changes. We're Nava, and today we’re coming out of stealth with an $8.3M seed round from @polychain, @archetypevc, @hack_vc, @FalconXGlobal, and @seedclub / @seedclubvc to build the trust infrastructure the agent economy is missing.
129
98
50
695
344,084
we'll be at @EFDevcon and @blockworksDAS! dm us if you'd like to meet ✨
5
4
1
24
1,488
speaking at @archetypevc’s annual general meeting in nyc today on the state of securing intelligence @vyas_krishnan @DannySursock @giovignone
5
2
1
29
1,638
The toolkit is growing and converging! What now becomes interesting isnt so much "can the agent do it" but "what do I want it doing" That's the one we'd add: a policy layer next to payments, so handing an agent a wallet stops being a leap of faith
All of these new agents are coalescing on baking in the same tools out-of-the-box: - Browser/computer use - Phone calls - Email + messaging - Calendar + scheduling - Durable memory/context - Payments/@link wallet - A credential store - Files/docs - Background jobs + monitoring - Bookings and IRL coordination If I were building the next Instinct/Muse/Grokbot, I'd add these by default, and also think about what other tools to weave in
7
2
34
1,568
Congrats @sfwallet on the beta launch! A wallet humans and agents use together, already running inside @ChatGPT Work and @claudeai Cowork. Excited to be working together on the verification layer: your agent proposes, Nava Guardian checks it against your policies, you approve knowing what it does. >> Get started with agents at superfluidwallet dot com
Introducing Superfluid Wallet: a wallet you and your agents use together. Now in beta testing. For on-chain finance and agentic commerce where agents act for people and people stay in charge. Understand what you’re authorising. Control what your agent can do. Let participating services verify who your agent represents. Built With the Best: @turnkeyhq , @proof , @circle , @navaai Read the launch article : xitter.twitterapi.xyz/sfwallet/status/210241…
9
8
40
1,904
My first trading agent placed a real ETH short before I could approve it. Nothing went wrong, but nothing would have stopped it either. 4 levels of financial autonomy, why a prompt isn't a mandate, and why limits on an agent have to live outside the model.
10
5
3
58
3,327
Explore what we're building with our our newly launched docs... → Nava Guardian: evaluate an action against your policy before execution → Nava Gateway: policy-aware x402 payment gating → NavaClaw: a DeFi agent harness to start from plug in wherever you want coming soon™️
15
45
1
91
3,733
We're working with @privy_io to enable AI agents to transact from real wallets, safely. Privy grants scoped wallet access. Nava verifies every transaction before it executes. Check out the demo + and post below to see how it works 🧵
35
12
4
127
13,328
Two layers, two jobs @privy's new scoped authorization grants an agent the authority to act on a wallet. Nava verifies whether each proposed use of that authority should actually be allowed. Access, then verification.
2
15
881
We read 20+ DeFi trading agent repos They've converged on the same 9-step pipeline 8 of those steps are commodity The 9th is where almost every public failure of the last year happened (full 🧵and post linked below)
14
8
80
3,590
The pipeline looks like this: trigger → data → features → LLM call → parse + size → risk gate → submit → log → exit Teams are pouring engineering into the LLM step Almost nobody is investing in the gate that decides whether the model was right.
2
10
640
what makes an agent trustworthy with capital? 💭
11
2
1
51
1,893
Nava’s guardian infrastructure was inspired by dragons ❣️🐉 @wwwwednesday
Nava AI Visual identity → Motion → Website design
7
4
45
2,515
We’re live at the agentic finance summit at @Microsoft today ! @vijaymichalik
18
3
70
3,186
Our Head of Product @vijaymichalik is speaking at @AgenticSummit on June 3, during #NYTechWeek ! 25% off tix with NAVAAI25 !
Network-issued trust signals, open registries, portable agent reputation, production-grade "know your agent" requirements. The infrastructure for verifying agents at scale is still being assembled. This panel examines what it actually takes. On stage: @marco_derossi, @MetaMask @OliverJaros_, @CMT_Digital @JamesJohnBeck, @ensdomains @vijaymichalik, @navaai Moderator: Jason Chaskin, @ethereumfndn June 3, NYC · agenticfinance.xyz
6
3
43
4,020
"the agent economy has an accountability gap, not a capability gap" @vyas_krishnan on what's actually missing in how financial agents touch user funds, and how we're working to close the gap @11AMdotclub
12
4
57
3,063
◢ concept ◣ AI more trustworthy. Blockchains more useful. The agentic economy actually deployable, and far more valuable. ◢ execution ◣ Nava doesn't constrict your agent. It gives you controls: ▸ set boundaries around execution ▸ wire up guardrails ▸ verify every step, intent ↦ settlement If you're an agent builder, we'd love to chat with you Private testnet waitlist is open ◆ navalabs dot ai ◆
31
3
1
73
3,441
live today with @11AMdotclub
8
2
1
56
3,321
thinking about agents
Thinking about agents
25
2
3
87
5,312
It has never been a better time to deploy and monetize an agent with every action verified before it settles
It has never been a better time to deploy and monetize an agent
12
57
3,283
The agent economy has a trust problem Every section here >> KYA, scoped delegation, verifiable provenance, stablecoin settlement >> points to the same conclusion: "Trust must be hardcoded into the architecture itself." We're building this !
13
6
97
6,798
We recently published our research behind Nava's core verification engine. It was accepted at the LAST-X Workshop at NDSS 2026 and introduces two things: 1) A new benchmark dataset for evaluating intent-transaction alignment in DeFi; 2) Nava Arbiter, our hybrid validation framework that sits between what an agent is told to do and what it actually executes onchain This research is the foundation of what we are building. Production performance + protocol coverage will evolve as we expand beyond our initial testnet, and we'll publish results as we go Check out the research + join the waitlist for our private testnet ↓↓↓
24
4
1
102
5,808