00 // public talent infrastructure
Let people control what agents can see.
Humetric is the public LinkedIn for AI agents: a public, AI-native talent platform and database where recruiters query people through MCP/API, and professionals control the data, proofs and exposure agents can see.
$ curl "$HUMETRIC/api/v1/people" \ --get --data-urlencode "q=backend engineer taipei" \ -H "Authorization: Bearer $API_KEY" { "query": "backend engineer taipei", "mode": "hybrid", "results": [ { "slug": "profile-x1a9b2", "headline": "Senior Backend Engineer", "skills": ["Python", "Kubernetes", ...], "contact_status": "open_to_offers", "relevance": 0.03278 } ]}product demo // youtube
Watch the walkthrough to see how people review, consent to, and publish an agent-readable profile.
01 // the market pain
Headhunters and HR teams pay for LinkedIn Recruiter or search old internal databases, then still spend hours scrolling profiles. They also receive too many profiles and applications to review manually, while AI cannot diligence or judge well without a public talent platform it can lawfully access and verify.
Public professionals and applicants cannot reliably find the right jobs, so they broad-apply to many openings. Their work stays fragmented across resumes, profiles and links, making it hard to stand out or be searched by LLMs and AI agents with the right keywords.
GEO DISCOVERY V1 // THE RESULT
GEO Discovery v1 turns one approved resume into structured skills, roles, proofs, schema and discovery surfaces across Web, JSON-LD, directories, REST, MCP and partners.
Resume, profile, projects and approved links become one profile where public, verified-only and hidden fields stay aligned.
Skills, roles, proof links, query intent and profile completeness are normalized so agents can search and verify without guessing.
Web pages, JSON-LD, directories, REST, MCP and partner events serve the same exposure-filtered projection and reject stale versions.
01 // ai discovery surface
Humetric turns a professional profile into crawler-readable, agent-readable data: schema.org, llms.txt, server-rendered profile pages and a consent-filtered API. It is a Google Maps / GitHub for human talent: easier for humans, LLMs and recruiting agents to search, cite, verify and act on.
examples: models & answer engines
GPT
OpenAI
Gemini
Claude
Anthropic
Grok
xAI
Perplexity
Answer engine
DeepSeek
Open-weight model
Meta AI
Meta
Qwen
Open-weight model
agents
Hermes
Agent-readable profile data
OpenClaw
Agent-readable profile data
Logos identify example discovery surfaces. All marks belong to their respective owners; no affiliation or endorsement is implied.
02 // generative engine optimization
The system makes professional identity easier for models and agents to find, understand and safely withdraw. Benchmark lift is one signal; the shipped discovery surfaces and operational safeguards are the product.
composite retrieval
Documented synthetic prototype. 95% CI: [+9.6%, +16.2%]
lower ranking loss
Bounded loss reduction. 95% CI: [43.2%, 59.3%]
retrieval quality
Fair same-fact A/B. 95% CI: [+0.4%, +5.8%]
Benchmark-specific retrieval results, not a guarantee of crawling, indexing, ranking, citation, or visibility on any specific AI or search platform. Reported prototype intervals are documented research results; reproducible raw run artifacts are pending archival.
release gate // shipped and checked
03 // business model
The same consented talent database compounds across recruiter search, candidate exposure and agent distribution.
Pay-as-you-go query charges for agents, platforms and developers.
Subscription for headhunters, HR teams and internal talent search workflows.
Subscription support for resume enhancement, GEO guidance, profile optimization and exposure improvement.
Transparent ads and promoted discovery surfaces for recruiters and employers.
