Manifesto
Humans should not be the procurement API.
Agents are moving from chat to action. But every useful action eventually hits the same wall: create an account, verify an email, add billing, find the API key, paste it somewhere safe, and prove it works.
That is not intelligence work. It is account choreography. It should be handled by a governed runtime, not a human tab wrangler.
Agent Acquire SaaS gives agents a way to acquire approved tools directly. They bring a capability request. The runtime brings company identity, browser control, inbox access, payment policy, secret storage, audit logs, and verification.
The operating principles
Software accounts are becoming infrastructure, not admin work.
Agents should ask for capabilities, not wait for humans to copy API keys.
Company identity, payment, inbox, and vault access should be delegated through policy.
Full autonomy still needs receipts, audit trails, spend limits, and revocation.
A working API smoke test matters more than a successful signup screen.
The end state
An agent says what capability it needs. The platform chooses an approved vendor, creates the account, pays within policy, stores credentials in the vault, runs a smoke test, and returns a working secret reference. Nobody manages accounts by hand unless policy says a human decision is required.
op://Agent Acquire SaaS/PostGrid/API Key
The agent receives a verified capability reference, not a raw card, password, or unmanaged API key.