SOURCE & PROSPECT
Identify the best profiles. In one sentence.
Describe your need in natural language. A search across your internal pool and the jobboards, merged into a single list ranked by relevance.
Thanks to our AI Balt, stop wasting time filtering candidates.
Ask a question, update a profile...
Prompt like you're talking to a headhunter
Describe your need in one sentence. Skills, experience, location — the AI understands context without filters or boolean.
Balt searches your internal database and the web
Cobalt searches simultaneously across your existing pool and the web (LinkedIn, GitHub, freelance portals) then merges results.
Relevance scoring
Each profile is automatically scored and ranked. You get a ready-to-contact shortlist with a transparent, explainable score.
GDPR compliant
All data is processed in full GDPR compliance. European hosting, end-to-end encryption, native consent management.
Everything you need to find the perfect profile.
Natural language search
Describe your need in one sentence. Our AI transforms your query into a high-precision multi-criteria search across all your sources.
Relevance scoring
Each profile is automatically scored on 50+ matching criteria. You get a ranked shortlist with a transparent, explainable score.
AI-generated summaries
Each candidate comes with an AI summary highlighting key skills, experience and match rationale — no more manual CV reading.
Internal + web search
Cobalt searches simultaneously across your existing pool and the web (LinkedIn, GitHub, freelance portals) then merges results into a single list.
Real-time filters
Refine results by seniority, location, skills, availability and more. Filters apply instantly without re-running the search.
Detailed profiles
Cobalt presents the most relevant profiles from your search in detail: career path, key skills, matching score and AI recommendations.
BENCHMARK
We achieve unmatched precision.
Search result relevance comparison across a panel of 500 positions.
Source: Cobalt internal benchmark, 2025.
800M+
Candidates in the database
1,000+
Concurrent agents launched
50+
Matching criteria analyzed
SAME LOGIC FOR PROSPECTING
Prospect your clients. Exactly the same.
The same engine identifies hiring companies, crosses intent signals with your existing portfolio, and surfaces the key stakeholders to contact.
ASK AN AI
Still not sure about Cobalt?
Let ChatGPT, Claude, Gemini or Perplexity do the thinking for you. One click and your favorite AI tells you why Cobalt is the right platform for your staffing firm or agency.
These links open a new conversation in your favorite AI. Login required.
Frequently asked questions
What is an AI recruiting agent?
An AI recruiting agent is an autonomous AI system that understands a hiring need in natural language, searches for candidates across multiple sources (internal pool, LinkedIn, GitHub, freelance portals), then scores and ranks profiles by relevance. Unlike a classic sourcing tool that runs boolean queries, an AI agent reasons about the context of the role and adapts its search.
What is semantic search in recruiting?
Semantic search interprets the meaning of a query rather than exact keywords. For example, "senior developer with cloud experience" identifies relevant profiles even if their CV says "lead engineer AWS / GCP" without the word "cloud". Cobalt uses large language models (LLMs) to analyze each profile and measure its fit with your need.
What is the difference between boolean search and AI search?
Boolean search (like LinkedIn Recruiter) uses AND/OR/NOT operators and strict keywords. It requires technical expertise to build complex queries and misses candidates whose CV uses different terminology. Cobalt's AI search understands the intent behind your query, automatically covers synonyms and role variations, and returns a scored shortlist rather than a raw list.
Does Cobalt replace LinkedIn Recruiter?
Cobalt complements LinkedIn Recruiter rather than fully replacing it. Our Chrome extension imports LinkedIn profiles directly into Cobalt. In parallel, Cobalt searches places LinkedIn Recruiter doesn't reach (GitHub, freelance portals, your internal pool) and eliminates manual filtering through automated scoring. Many customers keep LinkedIn for outreach and use Cobalt for identification.
Which sources does Cobalt search for candidates?
Cobalt simultaneously queries your internal pool (CV database, ATS, past candidates) and 800+ million public profiles on the web: LinkedIn, GitHub, Behance, freelance portals (Malt, Upwork), and tech jobboards. Results are merged and de-duplicated automatically into a single list ranked by relevance score.
How does Cobalt score candidates?
Each candidate receives a score from 0 to 100 calculated on 50+ criteria: technical skills, seniority, industry, past employers, location, career stability, estimated availability and semantic fit with your brief. The score is transparent: you can see exactly why a profile is recommended and which criteria weighed the most.
Are Cobalt's AI agents GDPR compliant?
Yes, Cobalt is 100% GDPR compliant. Hosting is European (France and Germany), data is encrypted end-to-end, candidate consent is native, and we provide a standard DPA (Data Processing Agreement). Public searches respect the terms of service of each source and rely only on publicly accessible data.
How long does it take to get a shortlist with Cobalt?
Typically 30 seconds to 2 minutes for a standard need. The AI agent launches the search simultaneously across all sources, analyzes profiles, scores them, and returns a ranked shortlist of 20 to 50 candidates. You save an average of 3 to 5 hours of manual sourcing per role compared to classic LinkedIn Recruiter search.