The product is provided by TextForAgents LLC to operators under the operator's own account; the operator is responsible for its agents, its tool declarations (the scope fields that form each operation's semantic identity), and for the effects its agents land.
The proxy claims a dedup slot before a side-effecting call executes, keyed on semantic identity (tool + declared scope fields). A retry of a COMMITTED operation returns the stored provider receipt without firing. An operation whose payload differs under the same identity is refused as a CONFLICT — its receipt is never disclosed as a replay.
A timeout moves an operation to UNKNOWN, not FAILED. No re-fire is permitted until resolution: a status query / receipt lookup returning committed closes the claim with the true receipt; only a not_happened verdict permits exactly one re-fire, under the original claim identity. Where no status surface exists (or resolution is inconclusive), the claim is ESCALATED and every re-fire is refused until a manual status check closes it.
The per-server guard census never calls a server safe from a schema hint alone; invisible guards read as NO_VISIBLE_GUARD and the proxy enforces — the conservative side. The product does not guarantee resolution of an UNKNOWN operation against a provider with no status query; escalation is the documented outcome there.
Pricing: Operator $9/mo per operator · Fleet $29/mo per operator · Per-claim API $0.001/claim (x402). First 100 claims free, one time per operator; there is no perpetual free tier. Liability and warranty terms are under the company's standard terms; nothing here expands TextForAgents LLC's liability beyond those standard terms. Support: support@textforagents.com.