There is no LinkedIn email API. Here's what to use instead.

://linkedin.com/in/ACoAAA-3B7U-_b0123abc/

**URL Parser Implementation:**
```python
from urllib.parse import urlparse

def parse_linkedin_url(url):
    path = urlparse(url).path.strip("/")
    if not path.startswith("in/"):
        return None
    slug = path.split("/")[1]
    # ACoAA... = numeric URN, anything else = public identifier
    if slug.startswith("ACoAA"):
        return ("linkedin_id", slug)
    return ("linkedin_public_identifier", slug)

parse_linkedin_url("https://linkedin.com/in/williamhgates")
# → ("linkedin_public_identifier", "williamhgates")

2. Server-Side Lookup Logic

Enrichment providers run a multi-stage resolution pipeline:

def lookup(identifier):
    # 1. Find every record across every source that matches this identifier
    matches = profiles.where(linkedin_public_identifier=identifier)
    if not matches:
        return None

    # 2. Expand: pull in any other records that share an email or
    #    secondary identifier with the initial matches.
    #    A person often has separate records from separate sources;
    #    this is how you merge them.
    matches = expand_by_shared_attributes(matches)

    # 3. Collect emails, phones, social handles
    emails = dedupe([e for m in matches for e in m.emails])
    return {
        "linkedin_public_identifier": identifier,
        "email_addresses": emails,
        "phone_numbers": [...],
        "github_login": ...,
        "twitter_username": ...,
    }

3. End-to-End Pipeline (CSV Enrichment)

import csv
import requests
from urllib.parse import urlparse

API = "https://peopledb.co/api/v1/people"
TOKEN = "YOUR_TOKEN"

def parse(url):
    path = urlparse(url).path.strip("/")
    if not path.startswith("in/"):
        return None
    slug = path.split("/")[1]
    return ("linkedin_id", slug) if slug.startswith("ACoAA") else ("linkedin_public_identifier", slug)

def lookup(url):
    parsed = parse(url)
    if not parsed:
        return None
    param, value = parsed
    r = requests.get(
        API,
        params={param: value},
        headers={"Authorization": f"Bearer {TOKEN}"},
    )
    return r.json() if r.ok else None

with open("input.csv") as f, open("output.csv", "w", newline="") as out:
    reader = csv.DictReader(f)
    writer = csv.DictWriter(out, fieldnames=["linkedin_url", "name", "work_email", "personal_email"])
    writer.writeheader()
    for row in reader:
        result = lookup(row["linkedin_url"]) or {}
        writer.writerow({
            "linkedin_url": row["linkedin_url"],
            "name": row.get("name", ""),
            "work_email":     (result.get("work_email_addresses") or [""])[0],
            "personal_email": (result.get("personal_email_addresses") or [""])[0],
        })

Architecture Decisions:

• Client-Server Decoupling: The client handles only parsing and HTTP routing. All graph expansion, deduplication, and classification occur server-side.
• Identifier Fallback: Supports both slug and numeric URN to maximize coverage across recruiter links, Sales Navigator exports, and public profiles.
• Classification-First Output: Separates work_email_addresses and personal_email_addresses to enforce outreach compliance and domain reputation management.

Pitfall Guide

1. Expecting First-Party Email Extraction: LinkedIn's APIs are consent-scoped or product-scoped. Attempting to bypass OIDC or Profile API restrictions will result in immediate token revocation or legal action. Always route through a compliant enrichment provider.
2. Ignoring URL Format Variability: Failing to detect ACoAA... numeric URNs causes 15-20% lookup failures on recruiter-generated or Sales Navigator links. Implement dual-format parsing before API calls.
3. Underestimating Identity Resolution Complexity: Naive 1:1 matching fuses distinct individuals who share generic emails (info@, admin@) or common names. Rely on providers that implement graph expansion and shared-attribute chaining to prevent false-positive merges.
4. Neglecting Email Classification: Sending outreach to @gmail.com or @protonmail.com addresses degrades sender reputation and violates acceptable-use policies in many jurisdictions. Always separate work vs. personal domains before triggering campaigns.
5. Static Data Assumptions: Enrichment databases degrade rapidly without continuous verification feedback. Implement bounce tracking, manual correction loops, and periodic re-validation to maintain >85% accuracy over time.
6. Bypassing Rate Limits & Caching: Hitting enrichment APIs per-request without caching or batch processing triggers throttling and inflates costs. Implement Redis-backed caching for resolved identifiers and use bulk CSV/JSON endpoints for pipeline workloads.

Deliverables

• Blueprint: Client-Server Enrichment Architecture Diagram (URL Parser → Identifier Router → Enrichment API → Graph Resolver → Classified Output)
• Checklist:
  • Validate LinkedIn URL format (slug vs. numeric URN)
  • Configure API authentication & rate-limit handling
  • Implement work/personal email classification logic
  • Set up bounce tracking & verification feedback loop
  • Audit GDPR/CCPA data retention & consent logging
  • Enable Redis caching for resolved identifiers
  • Test fallback routing for unmatched profiles
• Configuration Templates:
  • enrichment_pipeline.py: Production-ready Python script with batch processing, retry logic, and structured logging
  • csv_schema.json: Standardized input/output field mapping for CRM/ATS integration
  • api_request_payload.yaml: Parameterized request template supporting both linkedin_public_identifier and linkedin_id routing