What works today
Status, Updated
We are an early-stage startup. Working in lab means it runs end to end in our lab today, with a test business, tests and recorded evidence. In lab means it is on our build plan: designed, being built or tested. Nothing is live with real customers yet.
Working in lab runs end to end in our lab todayIn lab on our build plan: designed, being built or tested in our lab; not available yetWhat works today
Product by product
Every feature of every product, with where it stands today. Each product name opens its own page.
Discover
1 of 6 working in our lab
- One policy compiled into robots.txt, llms.txt, the AI catalog, agent cards and MCP server cards In lab
- Separate choices for AI search, AI training and AI agents, written the same way in every format In lab
- Every published file signed and checked for contradictions, and republished when your rules change In lab
- Agent card, commerce profile and sign-in details for your agent address Working in lab
- Which crawler read which file, and whether it then obeyed In lab
- A free Agent Readiness Report for your domain In lab
Shield
4 of 10 working in our lab
- Real agents verified by signature; fakes and replays refused Working in lab
- Customers give agents limited, expiring access instead of passwords Working in lab
- Protected pages and actions enforced at your server Working in lab
- Revoke a connector at once Working in lab
- Your rules per agent, action and amount, with nothing allowed by default In lab
- Identity from Web Bot Auth, Visa, Mastercard and other agent ID schemes In lab
- A session that read untrusted text and private data cannot act without a fresh yes In lab
- Limits on loops and retries per agent and per task In lab
- Checks at your CDN, and agents on your website flagged and sent to your front door In lab
- Pause every agent service with one switch In lab
Agent Storefront
4 of 9 working in our lab
- One official agent address for every assistant Working in lab
- Answers from reviewed facts and live systems, with source and freshness Working in lab
- Public, verified-agents-only and never: you choose what each fact may reach Working in lab
- Questions agents could not get answered become review tasks Working in lab
- Signed, expiring fact cards for prices, stock, hours and terms In lab
- A public check: did this business say this, and when? In lab
- Hidden instructions in reviews and feeds caught before they reach an answer In lab
- Changes sent to search engines and agents as they happen In lab
- Your agent address on your own domain, like agent.yourstore.com In lab
Connect
2 of 8 working in our lab
- Connectors that can never charge, pay, approve or confirm an order by themselves Working in lab
- Credentials kept as references, never shown to agents Working in lab
- Upload an OpenAPI file: each operation becomes read, hold, book, buy, change or never In lab
- Governed tools published to MCP, A2A, UCP and ACP at once In lab
- Signed tool lists: any change shows a diff and waits for your approval In lab
- Agent tokens never reach your systems; each call gets its own short-lived token In lab
- A test bench that shows what an agent can call and what your policy refuses In lab
- Shopify, booking and banking connectors with pilot partners In lab
Approve & Pay
5 of 8 working in our lab
- Nothing happens without the customer’s exact yes Working in lab
- A fresh yes whenever the price or terms change Working in lab
- Payment confirmed only by your payment provider Working in lab
- Card details and one-time codes never reach agents Working in lab
- One order per approval, even when 100 attempts arrive at once Working in lab
- Approval for transfers, claims, cancellations, plan changes and appointments In lab
- Approve from inside the assistant, on your own page, with a passkey In lab
- Real payment providers and agent checkout protocols In lab
Conversations
3 of 8 working in our lab
- Questions, updates and alternatives while a task is open Working in lab
- Consent with a purpose, a limit, an expiry and quiet hours, checked at every delivery Working in lab
- Follow-ups the customer agreed to, delivered to their agent Working in lab
- One case, whatever protocol the agent used In lab
- Each message sent the best way the customer’s assistant can receive it In lab
- Consent rules per channel and per country In lab
- Your staff take over a conversation in the same channel In lab
- Reminders before an approval or a hold expires In lab
Insights
2 of 6 working in our lab
- What agents asked for, and what you could not answer Working in lab
- Observed, stated, inferred and confirmed kept apart, with small groups hidden Working in lab
- The journey from question to approved order, by verified assistant In lab
- Verified versus claimed identity, and refusals by reason In lab
- Your AI search reports from Bing and Google joined to what agents asked you In lab
- Exports to Google Analytics and your warehouse In lab
Offers
1 of 4 working in our lab
- Labelled offers and discounts you fund, inside a task Working in lab
- Verified Offers: signed, expiring and labelled, checkable by anyone In lab
- Ads in AI assistant ad programs, built from reviewed facts, every campaign paused until you approve In lab
- Licensing and pay-per-access for content businesses In lab
The Proof Record
1 of 5 working in our lab
- Signed receipts for every order and refusal Working in lab
- A tamper-evident log of every decision and receipt In lab
- The Proof Record linking all five proofs In lab
- A public verifier that checks a record without access to our database In lab
- Dispute packs for payment providers and agent payment standards In lab
How we build the rest
In this order, because each stage depends on the one before it. We do not give dates: every stage starts only when the one before it has passed its tests.
Foundations
Key rotation and published key sets, a secret vault, one guarded exit for outbound calls, a public tamper-evident log, quotas, and each business walled off in the database.
The three engines
The Policy Engine at every entry point, session safety and loop limits, one case across protocols, the Proof Record, signed facts and the public verifier.
The first products
Discover and the free readiness report, Connect, answers built for AI search, funnel insights, CDN modules and agent identity schemes.
The full journey
Delivery to every assistant, consent rules per country, staff handoff, approval for every action type and the industry templates.
Real assistants and pilots
Real AI assistants connected, design partners, payment providers and your own domain, each switched on only with approval.
Growth and the network
Verified Offers, optional advertising, licensing, and the opt-in Trust Network after privacy review.
Questions about availability
What works today?
TrueSession is early. The core runs end to end in our lab with a test business: verified agents, answers from reviewed facts, customer approval, confirmed payment, one order per approval, updates during a task, consent checks and the dashboard. Anything marked In lab is on our build plan for the first pilots.
Is TrueSession available today?
TrueSession is in early development. Its core flows work in our synthetic merchant laboratory, and we are preparing the first pilots. Join the pilot to help shape the first deployments.
Help shape the first agent gateways
We are designing pilots with merchants, marketplaces and banks.