Loading…
Career Intelligence API
Give your AI access to live job retrieval, deterministic ranking, and explainable fit. Your model understands résumés; it can't see today's hiring market.
{
"title": "Senior iOS Engineer",
"score": 92,
"priority": "high_priority",
"matchExplanation": "Swift, AppKit, platform security",
"rejectionRisks": ["No Kubernetes exposure"]
}
POST /v1/match
Authorization: Bearer sk_live_a1b2c3…
{
"candidateSchemaVersion": 1,
"sourceText": "Senior backend engineer,
8 yrs. Python, Go, Kubernetes…",
"candidate": {
"skills": ["python", "go", "kubernetes"],
"locations": ["Remote"]
},
"limit": 25
}
{
"matches": [
{
"externalId": "job_8f2a3c…",
"company": "Stripe",
"title": "Staff Backend Engineer",
"location": "Remote · US",
"url": "https://stripe.com/jobs/…",
"score": 91,
"priority": "high_priority",
"skillMatchPct": 88,
"matchExplanation": "Strong overlap: Go,
distributed systems.",
"rejectionRisks": ["Listing emphasizes Kafka."],
"confidence": 0.82
}
// + 71 more, ranked by score
],
"corpusVersion": "2026-06-29",
"processing": { "latencyMs": 612, "jobsScanned": 48213 }
}
The decision
One résumé, two roles that both match on "engineer" and "Swift." Seeker scores them on evidence, not keyword overlap, and shows its reasoning.
Senior iOS Engineer
Evidence in the résumé
Backend Infrastructure Engineer
Keyword overlap (why a search ranks it high)
Rejected because
A generic LLM might recommend both. A keyword search definitely would. Seeker makes the call your users would trust, and hands your agent the reasoning to show them.
Live job retrieval
Every match is a real, currently-open listing, retrieved from a corpus that updates continuously. Nothing is generated.
Reproducible results
The same résumé returns the same ranking on every call. Deterministic scoring, not model sampling, so you can test it, cache it, and trust it.
Explainable fit scores
Each match carries a fit score, the skill gaps, and the reasoning, so your agent decides on evidence it can show the user.
You send a résumé
Plain text is enough, and the résumé text is what drives the match. No parsing or preprocessing required on your side.
Seeker scores the live corpus
The deterministic engine ranks active listings by skill overlap, seniority fit, domain relevance, and title intent.
Your agent gets evidence
Ranked matches with explanations come back as clean JSON. Your AI turns them into honest guidance for the user.
This is invite only. Keys are issued by hand after a quick review, so tell us what you are building. We read every submission and reach out when it fits. No timeline promises.