Overview
Profile Discovery takes a username from any supported platform — LinkedIn, GitHub, or Twitter/X — and finds the same person’s profiles on the other platforms. It combines index lookups, search engine results, and LLM-powered entity matching to return high-confidence matches.Input: A single username or profile URL for LinkedIn, GitHub, or Twitter/X
Output: Matching profiles across all three platforms, each with confidence score, profile context, and direct link
Confidence threshold: Only results scoring >= 0.45 are returned
Output: Matching profiles across all three platforms, each with confidence score, profile context, and direct link
Confidence threshold: Only results scoring >= 0.45 are returned
How It Works
- Profile lookup - The input username is looked up in our index (and live APIs where available) to get the person’s name, headline, and profile context
- Cross-platform resolution - Known linkages are resolved via reverse index lookups, embedded URLs in bios, and direct field mappings (e.g. GitHub’s
twitter_usernamefield) - Candidate search - For any platforms not yet resolved, Google SERP is queried to find candidate profiles
- Entity matching - An LLM scores each candidate on how likely it belongs to the same person, using profile context from both sides
- Filtering - Results below the 0.45 confidence threshold are dropped, organization/company accounts are excluded, and duplicates are removed (keeping the highest confidence per platform)
Quick Start
Sample Response
Starting from Any Platform
You can discover profiles starting from any of the three supported platforms:Use Cases
- Recruiting - Find a candidate’s GitHub to assess technical skills alongside their LinkedIn
- Sales intelligence - Build a fuller picture of prospects across platforms
- Identity resolution - Link profiles that belong to the same person across data sources
- Enrichment - Use the
contextfield to get a rich text summary of each discovered profile
Limitations
- Discovery relies on public search results, so profiles with very common names may produce fewer high-confidence matches
- The
contextfield may benullif profile data is unavailable for enrichment