HrFlow background
Most Accurate Look-Alikes Similarity for Profile ↔ Profile & Job ↔ Job

Matching API

The Matching API is a high-throughput similarity search engine that finds related Profiles for a Profile and related Jobs for a Job. Powered by Hiring Superintelligence, it uses HR-native embeddings, multilingual representation, and deterministic ranking to power deduplication, clustering, and look-alike talent discovery without taxonomies or keyword tuning.

Read the Docs
+98% accuracy
10M/s index throughput
1.2B trainset
GDPR & EU AI Act
99.99% uptime
output.json
{
  "score": 0.98,
  "profile": {
    "name": "Jane Doe",
    "similarity": "look-alike"
  }
}

Trusted by Customers, Partners & the AI Ecosystem

ManpowerLogo
CritLogo
EuropeanUnionLogo
SanofiLogo
EnhanceCVLogo
TeamtailorLogo
API REQUEST

Hybrid Similarity & Search

Query Profiles with an anchor profile or Jobs with an anchor job, and get ranked hits (JSON) with similarity scores plus pagination metadata (total, page, limit, …). Results are deterministic: same inputs → same order.

Select Matching Type

Query similar profiles → returns ranked Profile hits (JSON) + Similarity Scores (1-score, score) + pagination metadata (total, page, limit, …).

Explore

API GuideAPI EndpointTry Live Demo

Top Query Fields

Anchor Profile

key or reference, and source

Search Params

(Searching API): keywords, Geo, Ranges, Facets, Taxonomies, Raw filters

Score Threshold

minimum similarity score

Pagination & Sorting

index (source/board), page, limit, sort_by, order_by

response.json
 1{
 2  "code": 200,
 3  "message": "Matching completed successfully.",
 4  "meta": {
 5    "page": 1,
 6    "limit": 10,
 7    "total": 312
 8  },
 9  "data": [
10    {
11      "profile": {
12        "key": "abc123",
13        "info": {
14          "first_name": "Sarah",
15          "last_name": "Chen",
16          "location": "San Francisco, CA"
17        },
18        "skills": ["Python", "ML", "NLP"],
19        "experiences_duration": 7.2
20      },
21      "score": 0.97
22    },
23    {
24      "profile": {
25        "key": "def456",
26        "info": {
27          "first_name": "James",
28          "last_name": "Park",
29          "location": "New York, NY"
30        },
31        "skills": ["Data Science", "PyTorch"],
32        "experiences_duration": 5.1
33      },
34      "score": 0.91
35    }
36  ]
37}
CUSTOMER STORIES

Don't take our word for it!

Trusted by fast-growing HR Tech and Global Enterprise

“

Outcome-based similarity, not just cosine

vs. Cohere
Before HrFlow.ai

Cohere gave us cosine similarity, not outcome-based similarity. Two profiles could be equally close in vector space, yet only one consistently succeeds. We kept adding heuristics to compensate.

After HrFlow.ai

HrFlow.ai Profiles Matching replaced that with an HR-native Profile encoder trained on real hiring and application signals, so similarity reflects outcomes, not just semantics.

“

No more search-stack hacks for similarity

vs. ElasticSearch
Before HrFlow.ai

Elasticsearch wasn’t built for Profile→Profile or Job→Job similarity. We stitched together keyword overlap, boosting rules, and RRF, then spent months patching multilingual and seniority edge cases.

After HrFlow.ai

HrFlow.ai Matching gave us zero-config, production-ready deterministic similarity with hybrid filters, without the engineering headaches.

“

Cosine similarity is not HR similarity

vs. Pinecone & HuggingFace
Before HrFlow.ai

Pinecone solved storage and ANN retrieval, but not HR similarity quality. We kept cycling through Hugging Face encoders, managing migrations, and adding heuristics—yet results stayed inconsistent across roles and industries.

After HrFlow.ai

HrFlow.ai Matching shipped the full HR-native stack: Profile/Job encoders, deterministic scoring, custom features, and reasoning—so we stopped building the missing layers ourselves.

“

Outcome-based similarity, not just cosine

vs. Cohere
Before HrFlow.ai

Cohere gave us cosine similarity, not outcome-based similarity. Two profiles could be equally close in vector space, yet only one consistently succeeds. We kept adding heuristics to compensate.

After HrFlow.ai

HrFlow.ai Profiles Matching replaced that with an HR-native Profile encoder trained on real hiring and application signals, so similarity reflects outcomes, not just semantics.

“

No more search-stack hacks for similarity

vs. ElasticSearch
Before HrFlow.ai

Elasticsearch wasn’t built for Profile→Profile or Job→Job similarity. We stitched together keyword overlap, boosting rules, and RRF, then spent months patching multilingual and seniority edge cases.

After HrFlow.ai

HrFlow.ai Matching gave us zero-config, production-ready deterministic similarity with hybrid filters, without the engineering headaches.

“

Cosine similarity is not HR similarity

vs. Pinecone & HuggingFace
Before HrFlow.ai

Pinecone solved storage and ANN retrieval, but not HR similarity quality. We kept cycling through Hugging Face encoders, managing migrations, and adding heuristics—yet results stayed inconsistent across roles and industries.

After HrFlow.ai

HrFlow.ai Matching shipped the full HR-native stack: Profile/Job encoders, deterministic scoring, custom features, and reasoning—so we stopped building the missing layers ourselves.

“

Outcome-based similarity, not just cosine

vs. Cohere
Before HrFlow.ai

Cohere gave us cosine similarity, not outcome-based similarity. Two profiles could be equally close in vector space, yet only one consistently succeeds. We kept adding heuristics to compensate.

After HrFlow.ai

HrFlow.ai Profiles Matching replaced that with an HR-native Profile encoder trained on real hiring and application signals, so similarity reflects outcomes, not just semantics.

“

No more search-stack hacks for similarity

vs. ElasticSearch
Before HrFlow.ai

Elasticsearch wasn’t built for Profile→Profile or Job→Job similarity. We stitched together keyword overlap, boosting rules, and RRF, then spent months patching multilingual and seniority edge cases.

After HrFlow.ai

HrFlow.ai Matching gave us zero-config, production-ready deterministic similarity with hybrid filters, without the engineering headaches.

“

Cosine similarity is not HR similarity

vs. Pinecone & HuggingFace
Before HrFlow.ai

Pinecone solved storage and ANN retrieval, but not HR similarity quality. We kept cycling through Hugging Face encoders, managing migrations, and adding heuristics—yet results stayed inconsistent across roles and industries.

After HrFlow.ai

HrFlow.ai Matching shipped the full HR-native stack: Profile/Job encoders, deterministic scoring, custom features, and reasoning—so we stopped building the missing layers ourselves.

“

Outcome-based similarity, not just cosine

vs. Cohere
Before HrFlow.ai

Cohere gave us cosine similarity, not outcome-based similarity. Two profiles could be equally close in vector space, yet only one consistently succeeds. We kept adding heuristics to compensate.

After HrFlow.ai

HrFlow.ai Profiles Matching replaced that with an HR-native Profile encoder trained on real hiring and application signals, so similarity reflects outcomes, not just semantics.

“

No more search-stack hacks for similarity

vs. ElasticSearch
Before HrFlow.ai

Elasticsearch wasn’t built for Profile→Profile or Job→Job similarity. We stitched together keyword overlap, boosting rules, and RRF, then spent months patching multilingual and seniority edge cases.

After HrFlow.ai

HrFlow.ai Matching gave us zero-config, production-ready deterministic similarity with hybrid filters, without the engineering headaches.

“

Cosine similarity is not HR similarity

vs. Pinecone & HuggingFace
Before HrFlow.ai

Pinecone solved storage and ANN retrieval, but not HR similarity quality. We kept cycling through Hugging Face encoders, managing migrations, and adding heuristics—yet results stayed inconsistent across roles and industries.

After HrFlow.ai

HrFlow.ai Matching shipped the full HR-native stack: Profile/Job encoders, deterministic scoring, custom features, and reasoning—so we stopped building the missing layers ourselves.

“

Built-in fairness and compliance

vs. Open-Source Encoders
Before HrFlow.ai

With open-source encoders, we hit major fairness and compliance risks. A male anchor profile often returned mostly male similar profiles, creating allocation and representation biases we couldn’t justify under GDPR or the EU AI Act.

After HrFlow.ai

HrFlow.ai Matching gave us deterministic similarity with built-in fairness controls and HR-safe ingestion.

“

From semantic twins to hiring twins

vs. OpenAI
Before HrFlow.ai

OpenAI embeddings treated past and recent experience similarly, and missed certifications, seniority fit, and trajectory signals. We didn’t want semantic twins—we wanted hiring twins.

After HrFlow.ai

HrFlow.ai Matching delivered Profile→Profile similarity grounded in hiring outcomes, so look-alikes actually behaved like look-alikes in the pipeline.

“

From semantic-similar to intent-similar jobs

vs. Google
Before HrFlow.ai

Google vectors mostly clustered job descriptions. We needed next-apply similarity: jobs a candidate is likely to click and apply to after liking a role.

After HrFlow.ai

HrFlow.ai Jobs Matching uses Job encoders trained on application signals, so similar jobs became similar intent, and our engagement and time-to-fill improved without extra tuning.

“

Built-in fairness and compliance

vs. Open-Source Encoders
Before HrFlow.ai

With open-source encoders, we hit major fairness and compliance risks. A male anchor profile often returned mostly male similar profiles, creating allocation and representation biases we couldn’t justify under GDPR or the EU AI Act.

After HrFlow.ai

HrFlow.ai Matching gave us deterministic similarity with built-in fairness controls and HR-safe ingestion.

“

From semantic twins to hiring twins

vs. OpenAI
Before HrFlow.ai

OpenAI embeddings treated past and recent experience similarly, and missed certifications, seniority fit, and trajectory signals. We didn’t want semantic twins—we wanted hiring twins.

After HrFlow.ai

HrFlow.ai Matching delivered Profile→Profile similarity grounded in hiring outcomes, so look-alikes actually behaved like look-alikes in the pipeline.

“

From semantic-similar to intent-similar jobs

vs. Google
Before HrFlow.ai

Google vectors mostly clustered job descriptions. We needed next-apply similarity: jobs a candidate is likely to click and apply to after liking a role.

After HrFlow.ai

HrFlow.ai Jobs Matching uses Job encoders trained on application signals, so similar jobs became similar intent, and our engagement and time-to-fill improved without extra tuning.

“

Built-in fairness and compliance

vs. Open-Source Encoders
Before HrFlow.ai

With open-source encoders, we hit major fairness and compliance risks. A male anchor profile often returned mostly male similar profiles, creating allocation and representation biases we couldn’t justify under GDPR or the EU AI Act.

After HrFlow.ai

HrFlow.ai Matching gave us deterministic similarity with built-in fairness controls and HR-safe ingestion.

“

From semantic twins to hiring twins

vs. OpenAI
Before HrFlow.ai

OpenAI embeddings treated past and recent experience similarly, and missed certifications, seniority fit, and trajectory signals. We didn’t want semantic twins—we wanted hiring twins.

After HrFlow.ai

HrFlow.ai Matching delivered Profile→Profile similarity grounded in hiring outcomes, so look-alikes actually behaved like look-alikes in the pipeline.

“

From semantic-similar to intent-similar jobs

vs. Google
Before HrFlow.ai

Google vectors mostly clustered job descriptions. We needed next-apply similarity: jobs a candidate is likely to click and apply to after liking a role.

After HrFlow.ai

HrFlow.ai Jobs Matching uses Job encoders trained on application signals, so similar jobs became similar intent, and our engagement and time-to-fill improved without extra tuning.

“

Built-in fairness and compliance

vs. Open-Source Encoders
Before HrFlow.ai

With open-source encoders, we hit major fairness and compliance risks. A male anchor profile often returned mostly male similar profiles, creating allocation and representation biases we couldn’t justify under GDPR or the EU AI Act.

After HrFlow.ai

HrFlow.ai Matching gave us deterministic similarity with built-in fairness controls and HR-safe ingestion.

“

From semantic twins to hiring twins

vs. OpenAI
Before HrFlow.ai

OpenAI embeddings treated past and recent experience similarly, and missed certifications, seniority fit, and trajectory signals. We didn’t want semantic twins—we wanted hiring twins.

After HrFlow.ai

HrFlow.ai Matching delivered Profile→Profile similarity grounded in hiring outcomes, so look-alikes actually behaved like look-alikes in the pipeline.

“

From semantic-similar to intent-similar jobs

vs. Google
Before HrFlow.ai

Google vectors mostly clustered job descriptions. We needed next-apply similarity: jobs a candidate is likely to click and apply to after liking a role.

After HrFlow.ai

HrFlow.ai Jobs Matching uses Job encoders trained on application signals, so similar jobs became similar intent, and our engagement and time-to-fill improved without extra tuning.

🔗 INTEGRATIONS

Works with the tools you use

Integrate 200+ tools with the flip of a switch.

Integration 0
Integration 1
Integration 2
Integration 3
Integration 4
Integration 5
Integration 6
Integration 7
Integration 8
Integration 9
Integration 10
Integration 11
Integration 0
Integration 1
Integration 2
Integration 3
Integration 4
Integration 5
Integration 6
Integration 7
Integration 8
Integration 9
Integration 10
Integration 11
Integration 0
Integration 1
Integration 2
Integration 3
Integration 4
Integration 5
Integration 6
Integration 7
Integration 8
Integration 9
Integration 10
Integration 11
Integration 0
Integration 1
Integration 2
Integration 3
Integration 4
Integration 5
Integration 6
Integration 7
Integration 8
Integration 9
Integration 10
Integration 11
HrFlow.ai Data Studio

HrFlow.ai Data Studio

HR-native ETL with 200+ connectors plus Webhooks to ingest, normalise, and sync jobs & profiles across your stack—reliable pipelines with unified schemas.

Zapier ETL

Zapier ETL

No-code automation platform with 8,000+ app integrations to move data between tools using triggers + actions.

Make.com ETL

Make.com ETL

Visual automation platform to extract/transform/route data across 3,000+ apps (plus HTTP modules for any API).

Microsoft Flow ETL

Microsoft Flow ETL

Microsoft Power Automate—workflow automation with 1,000+ API connectors (and support for custom connectors).

Workato ETL

Workato ETL

Enterprise iPaaS/automation platform with 1,200+ pre-built connectors for orchestrating integrations and data workflows at scale.

Salesforce Flow Automation

Salesforce Flow Automation

Salesforce's low-code workflow automation tool; extended via AppExchange with 7,000+ apps to add integrations and capabilities.

🚀 KEY FEATURES

End-to-End Similarity Search for Large-Scale Matching, Explainability, and Responsible AI

HrFlow.ai Matching is a large-scale search engine for Profile-to-Profiles and Job-to-Jobs similarity search. It combines HR-native deep hierarchical encoders (full career trajectories, job requirements, context, and work environment) with deterministic similarity scoring and a two-stage pipeline (high-throughput retrieval/ANN index + optional refinement) that supports multilingual and cross-lingual representations. It supports hybrid search filters, explainability (Reasoning API), and feature fusion (tags, metadata, tabular signals)—powered by specialized Profile encoders and Job encoders trained with built-in fairness regularization and representation-bias calibration, aligned with the EU AI Act and international ethical AI requirements.

Tutorial Video3:45

🔒 ENTERPRISE-READY

Trust & Security

Built for sensitive HR data—secure by default, enterprise-ready.

01

Encryption

TLS in transit + encryption at rest to protect documents and extracted data.

02

Retention control

Minimal storage by default, with configurable retention policies to match your compliance needs.

03

AI-Act / GDPR / DPA ready

Built for sensitive HR data—secure by default, enterprise-ready. AI Act– and GDPR-ready processing, with documented controls for data handling and compliance.

04

Location / Region

Data processing and storage can be aligned with your required region (e.g., EU or US) depending on your deployment.

Visit Trust Center→
📊 FEATURE COMPARISON

HrFlow.ai Matching is the most accurate and production-ready Profiles↔Profiles and Jobs↔Jobs similarity engine

Feature
HrFlow.ai Matching
HrFlow.ai Matching
Cohere Vectors
Cohere Vectors
ElasticSearch
ElasticSearch
Pinecone DB
Pinecone DB
OpenAI Vectors
OpenAI Vectors
Deployment & Trust
Headquarters
🇫🇷 France
🇺🇸 USA
🇺🇸 USA
🇮🇱 Israel
🇺🇸 USA
🇺🇸 USA & 🇪🇺 EU Servers
Built-in
config
config
config
GDPR / AI-Act readiness
by design
config
HR Compliance (Safety & Guardrails)
Built-in
Pretraining Data
1,2B Hiring Signals
Noisy & Biased Web Data
Config (Depends on embedding)
Config (Depends on embedding)
Noisy & Biased Web Data
HR-Focused
Input Security (Prompt injection)
Pricing model
per record
per input tokens + record (expensive)
per Server (unpredictable)
per input tokens + record (expensive)
per input (expensive)
Vector Database
Built-in
Built-in
Built-in
Built-in
Max Index Storage
>2B records
>2B records
>2B records
>2B records
Refresh Speed (Index update / new record)
~1s
~1s
~1s
~1s
~5s
Speed (avg response / 1M records)
~2s
~2s
~1s
~2s
~14 days
DevOps burden (production scale)
Lowest
Medium
High
Medium
High
Deployment model
Managed API/Saas
Managed API/ Self-host
Managed API/ Self-host
Managed API/ Self-host
Managed API/ Saas
Core Technology
Technology
Deep hierarchical Encoders / Fairness & Bias Optimization
Deep Flat Encoders
Apache Lucene / TF-IDF encoder / Opensource Flat Encoders
Opensource Flat Encoders
Deep Flat Encoders
Multilingual
43 lang
23 lang
Config
Config
40 lang
Crosslingual
Match Scores
Outcome-based Similarity
Cosine Similarity
Cosine Similarity / Keywords Overlap / Reciprocal Rank Fusion (RRF)
Cosine Similarity
Cosine Similarity
Hiring Likelihood Profile similarity
Built-in (Profile encoder)
Application Likelihood Job similarity
Built-in (Job encoder)
White-collar Roles Accuracy
High
Low
Config (Depends on embedding)
Config (Depends on embedding)
Low
Blue-collar Roles Accuracy
High
Low
Config (Depends on embedding)
Config (Depends on embedding)
Low
Junior Roles Accuracy
High
Low
Config (Depends on embedding)
Config (Depends on embedding)
Low
Senior Roles Accuracy
High
Low
Config (Depends on embedding)
Config (Depends on embedding)
Low
Custom Feature Engineering
Built-in (Tags & Metadata)
Keyword boosting + Reciprocal Rank Fusion (RRF)
Fairness Regularization
Built-in (Constraints)
Data Calibration & Debiasing
Built-in (Pipeline)
HR Stack integrations (add-ons)
Hybrid Search
Built-in (Searching API)
Built-in
Config
Built-in
Reasoning & Explainability
Built-in (Reasoning API)
Built-in (matched keywords)
Resume, CV, Job parsers
Built-in (Parsing API)
Config
HR data enrichment & taxonomies
Built-in (Linking/Tagging/Asking APIs)
Jobboards / ATS / HCM / HRIS connectors
200+ connectors (Data Studio)
Candidate & Recruiter UI
Widgets (App Studio)
❓ COMMON QUESTIONS

Frequently Asked Questions

Everything you need to know about the Matching API

🧩 COMPLETE API SUITE

Go beyond the Matching API

Our APIs are designed to complement each other and unlock your data's full potential

Full Extraction API Suite

Transform HR documents into structured, enriched Talent & Workforce Data — powering every layer of Hiring Intelligence.

API Overview
94%82%71%58%

Ranking API Suite

Unlock Hiring Superintelligence at scale — with transparent, fair, and explainable ranking across every Talent signal.

API Overview

HrFlow.ai is an API-first company and the leading AI-powered HR data automation platform.

The company helps +1000 customers (HR software vendors, Staffing agencies, large employers, and headhunting firms) to thrive in a high-volume and high-frequency labor market.

The platform provides a complete and fully integrated suite of HR data processing products based on the analysis of hundreds of millions of career paths worldwide -- such as Parsing API, Tagging API, Embedding API, Searching API, Scoring API, and Upskilling API. It also offers a catalog of +200 connectors to build custom scenarios that can automate any business logic.

GDPR Compliant badgeCapterra 4.8 rating badge

SOLUTIONS

PRODUCT

RESOURCES

COMPANY