Customer Review Response & Risk Check

Draft a grounded response to one customer review, or withhold it when serious-risk signals need human handling. It never posts, contacts the reviewer, admits liability, or invents a remedy.

Capabilities and limits

Paste one customer review and supply the business name, rating, tone, and any verified resolution. ReviewReady returns a response draft, private internal summary, detected topics, offline-contact guidance, and an escalation assessment. It checks the supplied text for legal, safety, discrimination, payment, medical, employee, personal-information, violence, and self-harm concerns. Unsafe or unsupported public text is withheld. ReviewReady does not post, contact the reviewer, admit liability, or invent a refund, investigation, policy, promise, or completed action. Human review is required before posting. Model-backed requests use Dravara's configured OpenAI Responses API model with store set to false; applicable provider abuse-monitoring retention and Poe conversation retention may still apply. A validated ready-to-post response receives the approved Dravara creator charge. A withheld public response and a failed result do not receive the result charge. Correct validation errors before starting a new message. After a retryable failure, start one new message. If Poe's charge state is unclear, check its billing activity before retrying. Dravara keeps the exact paid result for replay for 365 days. After artifact deletion and final settlement, eligible operational identifiers are minimized; pseudonymous result-integrity, accounting, and settlement records remain under a separate recordkeeping period. The release verifies the approved Dravara creator charge and current Poe-displayed total in Rates before publication. Poe determines its separate platform component and may change the displayed total.

One run writes at most one result item. The Actor does not use Standby mode and does not perform actions outside the documented task.

Successful result event: review-response-created at $0.14.

Cost protection: model-backed requests accept at most 16 KiB of combined UTF-8 prompt and controls. Oversize requests do not call the model. See the live Pricing tab for current charges.