← All templates
Agent ToolSales & marketing
CRM Address Cleanup
A complete Agent Skill — copy it and drop it into your agent’s skills folder as crm-address-cleanup/SKILL.md and it can run this recipe on its own.
Turn a messy list of CRM or lead addresses into clean, canonical records — with an explicit review queue for anything ambiguous.
Bring-your-own address list
How it works
- Take the raw address list (a CRM export, a lead form dump, a spreadsheet) and call POST /v1/lookup once per row, in parallel.
- Split results by disposition: "resolved" rows get a clean canonical record (coordinate + jurisdiction); "clarify" rows carry multiple candidates for a human to pick between; "no_match" rows need manual research — including some that arrive as an HTTP error status rather than a 200 body, treated the same way, but any OTHER error status is a real failure, not a non-match.
- Write the clean list and the review-queue list back to the CRM as two separate buckets instead of one merged list of uncertain quality — never silently guess on an ambiguous row.
Example response
- clean
- disposition
- resolved
- resolved_address
- 480 BERDOLL LN, CEDAR CREEK, TX 78612
- lat
- 30.199699
- lng
- -97.496411
- confidence
- 0.95
- review_queue
- disposition
- no_match
- input
- 1100 King St W, Toronto
- reason
- address_too_coarse
- hint
- resolved only to place, which is not specific enough to identify a property.
- disposition
- no_match
- input
- not a real address
- reason
- address_too_coarse
- hint
- resolved only to place, which is not specific enough to identify a property.
Get a Mireye API key at mireye.com/account.