Job Talent Match API

Match every job against every talent, with context and not just keywords

Compare jobs and talent on 1 scale from 0 to 100, using the context in the data you send. Every pair comes back with the match score, the skills that matched and the skills that are missing. Close skills still count when the words are different.

No card needed. The call only reads the jobs and talent data you send.

1 API calljob × talent
Senior Platform Engineer Python · AWS · Troubleshooting · Distributed systems
Candidate A
AWSDebuggingKubernetes gap
94
Candidate B
PythonSystem designAWS gap
81
Candidate C
REST APIsDistributed systems gap
66
98.5%Match accuracy, measured on the matches already run
$0.0001What it costs to evaluate 1 job and talent pair
BillionsJob and talent pairs already scored on this engine
680,000Role defining skills in the context layer
180,000Job profiles, across 27 industries

The job is a senior engineer role and it asks for 6 skills. Each person it was compared with comes back with a match score out of 100, the skills that matched and the skills that are missing.

A software engineer
Remote, 60 months of experience
JOB-001 and TAL-001
Skills matched 5 of 6
software engineeringsystem designapi designtroubleshootingdata structures
Matched through a close skill 2
debugging covers troubleshootingsoftware development covers software engineering
Skills missing 1
ui frameworks
Experience matches the role Location matches Role title matches
A backend developer
Remote, 54 months of experience
JOB-001 and TAL-006
Skills matched 3 of 6
software engineeringapi designtroubleshooting
Matched through a close skill 2
debugging covers troubleshootingsoftware development covers software engineering
Skills missing 3
system designdata structuresui frameworks
Experience matches the role Location matches Role title does not match
A graduate developer
Remote, 10 months of experience
JOB-001 and TAL-009
Skills matched 2 of 6
software engineeringdata structures
Matched through a close skill 1
software development covers software engineering
Skills missing 4
system designapi designtroubleshootingui frameworks
Experience does not match the role Location matches Role title does not match
1 of 3
The problem today

Why matching fails

Most systems compare descriptions. Good matching has to understand what a person can actually do.

Matching on capability The job asks for troubleshooting The person wrote debugging 1 shared skill graph Both words point to the same skill, so the match is found and scored.
1

They match descriptions, not capability

Keywords and text similarity tell you what looks alike. They do not tell you whether the person can do the work the role needs.

2

The inputs are incomplete

Jobs leave requirements out. People describe their experience in their own words. Titles, skills and evidence change from 1 company and 1 system to the next.

3

There is no shared context

Without a model of how roles, skills and related capabilities connect, every match is 1 block of text compared against another.

Better matching needs more than better text comparison. It needs a shared model of capability.
Under the hood

How it works

OpenKnowra does not compare 2 documents. It reads the job and the person against the same skill graph, and compares them there.

Embeddings tell you what looks similar. A skill graph tells you how skills relate to each other.

The job The person Capability context Roles Tools and methods Skills Experience Domains Both sides are read against the same graph before they are compared.
1

Read the job

The requirements, the role context and the evidence in the job become a structured capability profile.

2

Read the person

Skills, experience and related evidence in the talent data are read against the same graph.

3

Compare in shared context

Both sides are compared against the same capability model instead of on raw text similarity.

The response

What you get back

Every pair in the call comes back with 1 score, the skills that matched and the skills that are missing.

A score out of 100

The same match score for every job and talent pair in the call, so results can be ranked straight away.

The skills that matched

See which job requirements the person's record supports, including the close skills that use different words.

The skills that are missing

See the gaps next to the score, so every answer carries a reason and not just a number.

Where it applies

Where it fits

1 matching capability, many workflows. The same call powers hiring, rediscovery, job recommendations, internal mobility, staffing, workforce planning and matching inside your own product.

Candidate discovery

Compare every applicant with the role and bring the stronger matches to the top.

Talent rediscovery

Score the people already in your database whenever a new role opens.

Job recommendations

Compare 1 person with many open jobs using the same match logic.

Internal mobility

Compare your own employees with internal roles before you go to the market.

Staffing and RPO

Match client demand against the talent pool you already have.

Bench and project staffing

Compare the people who are free with live project demand in 1 pass.

Workforce planning

Test the people you have against future roles before you decide to hire or build.

Job boards and marketplaces

Put job to talent matching inside the product your users already use.

Reason to believe

The context layer

Skills are not isolated labels. They connect to roles, related skills, tools, domains and experience, and that is what both sides of every match are read against.

The context foundation

An enterprise context layer that keeps growing, held together by a skill context graph

Both sides of every match are read against the same layer, which is why a job written 1 way and a profile written another way still end up speaking 1 skill vocabulary.

match accuracy
680,000Role defining skills
180,000Unique job profiles
27 industriesCovered by the layer
12 yearsCleaned up and grown at scale

Relationships, not just labels

The graph holds how skills and roles connect, so related skills and role context are part of the match.

Evidence becomes context

Skills, experience and role history add up to a capability profile, even when the person used different words for it.

1 frame for every comparison

The job and the person are read against the same context before the score is worked out, so scores stay comparable.

Pricing

You pay for 3 things, and only for what the call actually processed. Pay as you go, with no commitment.

Job context
$0.15
1.5 credits

Charged once for each job you send with a document.

Talent context
$0.20
2 credits

Charged once for each person you send with a document.

The pairs
$0.10
1 credit per 1000 pairs

Charged per 1000 pairs evaluated, rounded up. That is $0.0001 a pair.

Skills without a document are freeA job or a person sent as structured skills costs nothing on that side.
Failed calls are freeA call that does not produce a result is never billed.

10 jobs and 10 people, all with documents, is 100 pairs and costs $3.60, which is 36 credits.

4 MB per document 500 documents and 512 MB in 1 request pdf · doc · docx · html · txt

Read the full API reference →

See what your jobs and talent match today

10 jobs and 10 talents is 100 scored answers for 36 credits. Generate a key and run the call in the next 10 minutes.

Generate API Key