Recooty: AI ATS for Small Recruiting Teams
On this page 6 sections
Short answer: Recooty is an applicant tracking system for small recruiting teams and agencies that combines job posting, career pages, candidate pipelines, communication, scheduling, team collaboration, reporting, integrations, and AI-assisted screening. Its accessible packaging and broad feature list do not establish ranking accuracy or hiring impact. Buyers should test candidate data, AI outputs, plan limits, permissions, and integrations with their own roles.
Recooty’s current pricing page lists tiered plans, a trial, job and user allowances, candidate management, career pages, job distribution, scheduling, reports, SSO on higher tiers, and other features. It also publishes customer counts and brand logos. Those are vendor claims, not independently verified adoption evidence. This review uses product material available on September 13, 2026 and removes unsupported revenue and founder narratives from the earlier article.
Product boundary and fit
Recooty can organize the recruiting path from job creation through application, screening, interview, offer-related work, and reporting. It is not a payroll or global employee system of record. A buyer should specify where approved requisitions originate and where accepted hires go next.
The platform is most plausible for a small team that wants more structure than email and spreadsheets without a long enterprise deployment. It may also fit a recruiting agency with straightforward client and candidate processes. Fit should be determined by active jobs, applicant volume, team permissions, countries, required integrations, data retention, and review complexity rather than the label “SMB.”
Create a package matrix from the current quote:
| Requirement | Contract detail |
|---|---|
| Jobs, users, and candidates | Included limits, archive rules, overages |
| Job boards | Supported channels, approval time, paid media, source attribution |
| Careers page | Domain, language, accessibility, analytics, consent text |
| Candidate workflow | Stages, tasks, messages, interview and offer functions |
| AI | Included uses, input fields, output, explanation, human override |
| Security | Roles, SSO, logs, export, deletion, support access |
| Integrations | API entitlement, credentials, rate behavior, connector owner |
Do not assume every feature shown on a product page is included in a lower plan or available in every country.
AI ranking is a recommendation, not verified truth
Recooty’s AI candidate ranking page describes filters, matching scores, summaries, ranking, and automated notifications. It also advertises precision and time-saving figures. Those figures are Recooty marketing claims; the page does not establish independent validity for a buyer’s jobs.
The application-screening documentation shows that recruiters can filter by skills, experience, education, stage, and AI score and can see an AI summary with the candidate record. This clarifies the interface but not the model’s error rate or job relevance.
Evaluate each function separately:
- Resume parsing converts a document into structured fields and should be checked against the source resume.
- Matching compares a job description with candidate information and should expose the criteria used.
- Summaries compress source material and can omit or misstate relevant facts.
- Ranking orders people and can influence review opportunity even when it does not auto-reject.
- Knockout questions execute explicit rules and need an exception and accommodation path.
Use the voluntary NIST AI Risk Management Framework to record the use context, input, version, result, reviewer, override, and observed error. Test career gaps, nonstandard titles, different resume formats, multilingual records, incomplete data, and candidates with equivalent experience described in different terms.
Any score or automated rule used in a U.S. employment decision should be reviewed against the Uniform Guidelines on Employee Selection Procedures and applicable state and local law. Human review is necessary but does not make an invalid criterion valid.
Job distribution requires channel evidence
Recooty markets broad job-board distribution. The actual result depends on each board’s feed rules, employer verification, job content, geography, paid placement, and review process. A posted status in the ATS is not proof that the listing is live or indexed.
Test publish, update, pause, close, apply, and source attribution for each required channel. Compare started and completed applications, qualified review, interview, offer, and hire by source. Remove duplicate and spam applications before interpreting conversion. Vendor board counts and potential reach should not be used as hiring outcomes.
The candidate form should be tested on mobile, low bandwidth, keyboard navigation, and assistive technology. Collect only fields needed at that stage, show the employer and privacy notice clearly, and preserve the version of questions shown to each person.
API access reaches sensitive records
Recooty’s current API guide distinguishes a public-job API key from personal access tokens that can read candidate data or write records. It describes plan gates, permissions, token expiry, revocation, candidate endpoints, webhooks, and some beta write endpoints. This is useful implementation evidence and also a warning against placing access tokens in browser code.
Use one named credential per integration, store it outside source code, grant only required permissions, monitor use, and rotate before expiry. Define idempotency and duplicate handling for job and application writes. A beta endpoint should have a rollback and reconciliation process. Candidate resumes, screening answers, and AI insights copied into a warehouse or another system inherit separate access and deletion obligations.
Privacy and subprocessors need review
Recooty’s privacy policy identifies Recooty Inc. and describes services, candidate information, security measures, third-party processing, and rights. Its subprocessor list identifies group and cloud providers. These are vendor disclosures, not a conclusion that a particular employer’s use is compliant.
Map where candidate data is collected, hosted, enriched, messaged, exported, backed up, and deleted. Define lawful basis, notice, retention by candidate stage, access and correction, suppression from future outreach, and propagation to connected systems. Confirm the contract, data processing terms, transfer mechanism, incident notice, and deletion evidence for the employer’s jurisdictions.
Decision standard
Pilot Recooty with two different job families, an imported candidate, an inbound applicant, a duplicate, a withdrawn person, an AI ranking override, a failed job-board post, and an API-created record. Measure verified workflow time, parsing errors, ranking disagreements, missing messages, duplicates, source accuracy, integration failures, and admin effort.
Recooty is a reasonable option when a small team needs an ATS quickly and can operate within the contracted limits. It is a weaker fit when the buyer needs complex multinational configuration, deep employee lifecycle functions, or independently validated automated assessment. The purchase case should use a tenant-level pilot, not unsourced revenue, customer, precision, or productivity numbers.