Make it Proofable

Billions of AI agents are coming. Almost none of them can prove who they are, who they work for, or what they actually did.

Today we're introducing Proofable, the portable proof layer for people, businesses and AI agents.

Prove once, reuse everywhere, never restart from zero.

Trust resets every time you change systems

Every app asks the same questions. Are you a real person? Do you own this domain? Is this wallet clean? Who authorized this agent, and what was it allowed to touch?

Then it throws the answer away. The next app asks again. And when an agent's session ends, the record of what it did ends with it.

That was fine when AI answered questions in a chat box. It breaks the moment an agent can push code, move money or read customer data.

"Trust me" doesn't scale.

How it works

Proofable turns every check into a signed proof you own. Apps, gates and agents read the proof instead of running the check again.

Prove who you are. Proof of human through ZKPassport, without handing anyone your passport. Link X, GitHub, LinkedIn, Discord, Telegram and more to one profile.

Prove what you control. Domains, Google and Microsoft work accounts, wallets, tokens, NFTs and smart contracts.

Give every agent its own identity. Each agent is tied to the person or business behind it, with the actions, spend limits and expiry its owner signs off on.

Let context travel with the agent. Memory, rules, instructions and skills are stored as proofs, so an agent picks up where it left off in whatever tool you open next.

Run work that leaves a receipt. Agents run tasks now, on a schedule, or on a hosted computer. Sensitive steps pause for the owner's approval. Finished work leaves a signed record of what happened.

Gate anything. Turn any mix of proofs into an access gate for a community, a listing or an API.

Sell what you build. List agents, tools and data.

It runs inside the AI tools you already use over MCP (https://mcp.proofable.me/mcp) and in your own app with the API/SDK.

What our partners say

"Together, we are creating verifiable evidence from silicon to action, giving production AI systems a stronger foundation for trust." Phala

"Strong risk intelligence should compound across the ecosystem, not restart at every gate." Webacy

Yep, we have proof

• Top score on AgentGovBench, a perfect 48/48.

• 42.4x fewer tokens and 8.8x faster than searching raw records.

• 3.3x cheaper agent handoffs than redoing the work, and 3 of 3 tampered handoffs caught.

Every figure is published: https://docs.proofable.me/evidence

Open, for everyone

• CAIP-380 Portable Proof, the open standard.

• Built with Phala, ZKPassport, Webacy and context.dev. Smart contracts audited by SafeStackAI.

• NVIDIA Inception and Google for Startups program participants.

• Evidence used in the Linux Foundation AAIF Identity & Trust group.

Start

Free with 500 credits a month. Pro is $29.99 a month for 15,000 credits.

Start free: https://proofable.me/profile

Onboard your agent: https://docs.proofable.me/mcp/setup

We set up your first agents and first workflow and add 100,000 credits: https://proofable.me/pilot

To everyone who backed this before it could be explained, thank you.

Make it Proofable.