Let members search in plain language across the MLSs they’re authorized to access. Results include listing attribution, source information, and update times, with cross-listed duplicates merged.

Set the boundaries for data access, field visibility, attribution, and consumer sharing.Recursive AI applies those rules to answers, with brokerage and office overrides that cantighten your MLS’s requirements

Review the identity, filters, server calls, rules, and model behind a request. See which fields were sent to the model, which were withheld, and why an answer was redacted or blocked.

Members can compare listings across their connected MLSs in one request. Each source’s access scope and display rules remain in place.

Listing cards carry the listing office, MLS source, and update time. Members can openlisting details and inspect how the answer was produced.

Access Recursive AI on Mac, Windows, the web, and phones. Members can review listings and answer details at their desk or on the go.

Private remarks, showing instructions, and lockbox codes are withheld from model context. The request log identifies the fields that were included and those that were removed.

MLS data is not used for model training. Recursive AI uses zero data retention for the listing data processed to produce answers, while maintaining a separate MLS-owned audit trail.

Human approval required with every written change. Bulk downloads are blocked, and hidden instructions are ignored.

Your MLS server enforces which records a member can reach. Set boundaries for listing status, property type, geography, and sold-history lookback within the applicable data license.

Enforce IDX, VOW, and BBO scope alongside brokerage attribution, MLS logos, disclaimers, and update times. Apply seller opt-outs, state-specific sold-data rules, and AI photo labeling requirements

Your MLS establishes the rules. Brokerages and offices can add tighter restrictions, while role and audience permissions determine what each user can access or share.

Show members what changes when a listing goes to a consumer. Restricted listings are locked, confidential fields are removed, and seller display preferences are applied before sharing.

Include the required attribution, logos, disclaimers, and update times for each source. Share expiration follows the applicable MLS policy.

Members can edit the AI-drafted note before sending. Fair housing language checks help identify wording that needs revision.

Review the member’s identity, selected filters, server calls, applied rules, model, andtiming in one record.

See the names of fields provided to the answer model and the confidential fields that were withheld.

Tamper-evident, hash-chained entries preserve the request history. The log stores record IDs rather than copies of listing records and belongs to your MLS.