Find similar profiles using vector embedding similarity (pgvector).

Authentication required: Header api_key: <your_api_key>

Pricing: 0.25 credit per profile returned (DB data)

This endpoint uses OpenAI embeddings + pgvector cosine similarity to find profiles that are semantically similar to a given LinkedIn profile.

How it works:

1. Takes a LinkedIn profile URL, public_id, or member_identity
2. If the profile has no embedding yet: scrapes it live, generates embedding, stores it
3. Queries pgvector for the most similar profiles by cosine distance
4. Returns profiles ranked by similarity score (1.0 = identical, 0.0 = unrelated)

Use cases:

• sales: Find similar prospects (matches on title, industry, company size, skills)
• recruiter: Find similar candidates (matches on skills, education, experience)

Example request:

{
    "profile": "https://www.linkedin.com/in/johndoe",
    "use_case": "sales",
    "country": "US",
    "count": 25,
    "offset": 0
}

Example response:

{
    "lookalike_profile": "https://www.linkedin.com/in/john-doe",
    "total": 847,
    "offset": 0,
    "count": 25,
    "has_next_page": true,
    "credits_used": 6.25,
    "method": "vector",
    "results": [
        {
            "similarity_score": 0.8234,
            "member_identity": "ACoAABxxxxxx",
            "headline": "Sales Director at BigCo",
            ...
        }
    ]
}