Moonlit lets AI assistants earn research rewards without inbox access

Moonlit, from Vulture Labs, is launching a paid research model that lets people approve AI-drafted answers before sharing limited source details with buyers. The setup is designed to keep inbox access private while still paying participants for accepted answers.

Why it matters: - Moonlit aims to let people earn money from AI-assisted research without handing a research company access to an entire inbox. - The model keeps the person in control of what gets shared, which matters for privacy, security and sensitive email content. - Research buyers get usable answers and aggregated results, but not inbox connections or contact lists.

What happened: - Moonlit, a paid research service from Vulture Labs, says a participant's own assistant prepares answers to company questions. - The assistant can use permissions the person has already granted, such as finding an order confirmation. - Moonlit does not receive the assistant's mailbox connection. - Before submission, the assistant shows the person the proposed answer and the reward. - If the person approves, Moonlit can submit limited source details tied to the answer.

The details: - Shared source information can include a merchant domain, date, amount and a hashed message identifier. - In some cases, Moonlit may request the original email headers for verification. - Those headers are used for verification and then discarded. - Moonlit does not routinely require the full message body. - The privacy page outlines the information collected, the service providers involved and account controls. - Moonlit withholds groups smaller than five in buyer-facing results. - The service pays for accepted answers subject to verification and study availability. - Participation is limited to U.S. adults 18 and over. - Ready balances of at least $20 can be paid through Stripe after a seven-day verification hold. - Moonlit says companies can request a study at More information. - People can connect an assistant at the setup guide. - The privacy page is available at the privacy details.

Between the lines: - Moonlit is positioning AI as the worker, but the person as the gatekeeper. - That structure lowers the privacy tradeoff compared with models that require broad inbox access. - The design also suggests Moonlit wants answers that are easier to verify without collecting more data than needed. - The offer may appeal most to people already using an assistant for search and drafting, since the assistant can do the work while the person decides what is worth selling. - Sensitive questions are sent back to the person instead of being answered by inference.

What's next: - Accepted answers will continue to earn rewards when they pass verification and fit available studies. - People who already use an AI assistant can connect it through Moonlit's setup flow. - People new to AI can use the same setup guide to choose a supported assistant and prepare for paid surveys. - Earnings and work depend on study availability, eligibility and accepted answers.

The bottom line: - Moonlit is trying to turn AI-assisted research into paid work without demanding full inbox access, making privacy the main product feature.