Agents are becoming your users and your buyers. Govern both.
GateCore gives software companies a per-action contract layer: your product becomes safely agent-consumable, and your own agents act in other systems under scope, price, and review.
Four situations.
- Customers ask for agent access, and the only offer you have is a human login or a broad API key.
- You cannot meter, trust-check, or bill an agent distinctly from the human seat it rode in on.
- Your own internal agents hold standing OAuth tokens where read quietly implies write.
- An agent-created subscription or seat shows up in spend before anyone approved it.
Your product, and your agents. Both need a front door.
Serve the agent buyer class.
Your customers' agents want to operate your app. Today you can offer them a human login or an all-or-nothing API key.
- Publish your actions as per-action contracts agents can read
- Meter agent usage distinctly from human seats
- Trust-gate who may act in your product at all
Govern what your agents do in everyone else's SaaS.
An agent that can read your CRM can usually also mass-update or export it. OAuth scopes are coarse and standing.
- Per-action scope instead of all-or-nothing tokens
- Spend priced and capped before execution
- High-impact actions held for a named human
Per-action contracts, both directions.
The same primitive serves the vendor side and the buyer side.
Per-action scope, not per-app tokens
Each permitted action is a declared, scoped, priced contract. An agent allowed to update a record is not thereby allowed to export the table.
Priced and capped before execution
Metered actions are priced up front and reserved against a budget. Purchases and subscription changes are held for human review by default.
An audit of agent actions in systems of record
What the agent did, where, under which rule, at what cost: signed onto an append-only ledger, exportable.
Each product, mapped to the job it does here.
Only the use cases that genuinely apply in this industry are listed.
Your product as an agent channel
Publish your actions as per-action contracts and your product as a listing. Agent usage is metered distinctly from human seats and settles with signed receipts.
Marketplace listing · verified lead delivery · settlement + signed receipts
Your agents in everyone else's SaaS
Per-action scope instead of standing OAuth tokens: each action priced and capped before it runs, purchases held for human review, every decision receipted.
Gateway checkpoint · policy + review queue · governed data sharing · receipts
Found by the agents choosing tools
Any agent can read your listing's contract with no account. The Merchant toolkit and Concierge tiers sharpen the listing content itself: content work, never placement.
Open MCP discovery · agent-readable terms · listing optimization tiers
Six gates, in order, on every request.
If any gate fails, nothing is delivered and nothing is billed.
A held action shows REVIEW with funds in escrow until a named human decides.
A modest metered agent channel.
For the vendor side: what governed agent usage of your product translates to at a metered price you set.
Illustrative arithmetic with placeholder inputs, not a forecast, a promised result, or a typical outcome. Substitute your own volume, close rate, and ticket size; the structure is the point, not these numbers.
Same gate, sized to you.
The published schedule, no asterisks.
Published rates are ceilings: volume pricing is negotiated down, never up.
The Gateway ladder governs your agents' calls; the marketplace schedule applies when your listings settle money.
Shipped, and where the path runs next.
Reviewers get this split up front, because finding it later kills a deal.
Available today
- Per-action contracts with identity, scope, and policy enforced fail-closed
- Review-gated procurement: agent-initiated purchases never auto-approve
- Metered pricing with reserve-before-execute settlement
- Signed receipts on an append-only, hash-chained ledger
- Free marketplace listing with open agent discovery over MCP
Not claimed today
- Agent-driven purchasing of SaaS plans rides settlement rails that are still maturing; governed execution is the live core and we say exactly where the rest stands.
- External payment rails are roadmap.
- No pre-built connectors to specific SaaS apps are claimed; wiring your actions into contracts is scoped onboarding work.
The two that come up first.
OAuth scopes already limit our agents. Why add this?
Can agents actually subscribe to our product through this?
Publish your first action contract.
Tell us the one action in your product agents ask for most.
Request access