Score off-market real estate deals from emails
A pipeline has a Python worker read deal emails, Claude extract fields into fixed JSON, and code score them; the poster is stuck on incomplete data and login-walled sources. Task: Suggested implementation: Connect a dedicated Gmail inbox to a Python worker that splits multi-deal emails. Use a low-cost Claude model to extract fixed JSON fields, leaving unknowns null. Implement plain Python logic for single-family and multifamily scoring, triggering paid data lookups only after a basic price check passes. Human role: Review extracted data and approve final deal scores.; tools: Claude Code, Claude. Human role: The original post does not say how a person stayed involved.; limitations: Prerequisites include access to a dedicated email inbox and API keys for Claude and data lookup services. Cost control is critical; expensive lookups must be gated by cheap initial checks. Source reliability is low due to HTTP 403 error. Human approval is required before publishing or spending on data..
Tools
Published on Agent Work Atlas:
Overview
A pipeline has a Python worker read deal emails, Claude extract fields into fixed JSON, and code score them; the poster is stuck on incomplete data and login-walled sources.
Workflow
Suggested implementation: Connect a dedicated Gmail inbox to a Python worker that splits multi-deal emails. Use a low-cost Claude model to extract fixed JSON fields, leaving unknowns null. Implement plain Python logic for single-family and multifamily scoring, triggering paid data lookups only after a basic price check passes. Human role: Review extracted data and approve final deal scores.
Reported result
Actual outcomes are unknown as the source is inaccessible. Proposed acceptance check: Measure the percentage of leads successfully scored without manual intervention and track the rate of missing street addresses or unit counts in extracted JSON.
Human role
The original post does not say how a person stayed involved.
Limitations
Prerequisites include access to a dedicated email inbox and API keys for Claude and data lookup services. Cost control is critical; expensive lookups must be gated by cheap initial checks. Source reliability is low due to HTTP 403 error. Human approval is required before publishing or spending on data.
Try it
I need to set up an automated pipeline for scoring off-market real estate deals from emails. First, connect my dedicated Gmail [replaceable input: gmail_address] to a Python script. The script should split multi-deal emails into individual records. For each record, use the Claude API [replaceable input: api_key] to extract these fields into JSON: street_address, unit_count, asking_price, property_type. Leave unknown fields as null. Next, implement Python scoring logic: if asking_price is below [replaceable input: threshold], trigger a paid data lookup [replaceable input: lookup_service_api_key] to verify value. Do not run paid lookups if the initial price check fails. Finally, output a CSV report [replaceable input: output_path] with scores. Before running on live data, perform a dry run on 5 sample emails [replaceable input: sample_emails] and show me the JSON output for confirmation. If any critical fields are missing in the samples, ask me how to handle them before proceeding.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.