Findr
Findr is a Spanish AI recruiting platform whose agent, Marvin, runs the entire top of the hiring funnel: sourcing across many public channels, scoring against your rubric, personalised outreach, first-round screening and interview booking, handing recruiters only briefed finalists.
What is Findr?
Findr is a Madrid-based recruiting company whose product is Marvin, an AI agent it markets as "the AI recruiter that never sleeps". Where most sourcing tools stop at search filters, Marvin is designed to run the whole top of the hiring funnel by itself, continuously, and to hand a human team only the candidates worth meeting. The workflow starts with a brief. A recruiter pastes a job description or describes the role in plain language, and the agent asks the clarifying questions a senior recruiter would ask. From there Marvin works through seven named modules. Sourcing scans public signal well beyond LinkedIn — GitHub, Stack Overflow, X, Glassdoor, Dribbble, niche forums and conference archives — from a pool the company describes as more than five billion public profiles across twelve or more channels. Scoring ranks candidates against a rubric the customer defines, across roughly ten dimensions, and shows the reasoning behind each score. Screening runs a short voice or chat conversation and returns a transcript, a summary and a recommendation. Scheduling books interviews on connected calendars and handles the back-and-forth. Handoff produces a one-page interview brief for the hiring manager. Guardrails keeps the agent inside an inspectable playbook per role, with approval required before sending, competitor exclusions and bias checks. Integrations connect it to existing ATS, email, calendar and chat tools. Findr places unusual emphasis on bias prevention, claiming five architectural guarantees: identifying fields are stripped before any model reads a profile; a constitution overrides every prompt; several independent agents evaluate in parallel; everything is logged with its full context; and models run deterministically so the same profile yields the same evaluation. The company reports figures such as 81% less time to a first qualified candidate and a 73% reply rate, though it publishes no methodology. It names Allianz, Cabify, EY, Fever, Inditex, Acciona, IBM and Serveo as customers. The product is sold, not self-served: no price appears anywhere, and every call to action leads to a demo or a sales conversation.
What it does
- Source candidates across LinkedIn, GitHub, Stack Overflow, X, Glassdoor, Dribbble, niche forums and conference archives
- Score every candidate against a rubric you define, exposing the reasoning rather than just a number
- Write and send personalised outreach drawn from each candidate's public work, in your brand voice
- Run a first screening conversation by voice or chat, then summarise it with a recommendation
- Book interviews on connected calendars, handle reschedules and chase confirmations
- Hand the hiring manager a one-page brief covering fit, what to probe, red flags and expected compensation
- Rank your own employees against a role first, through internal mobility and skills mapping
When to use Findr / When not to
A quick filter to help you decide if Findr is the right fit.
When to use Findr
- In-house talent acquisition teams running many open roles at once and drowning in manual sourcing
- Mid-market and SMB companies that cannot justify a dedicated sourcing function but still hire steadily
- Recruitment agencies and RPO providers who need per-client workspaces and white-label outreach
- Technical recruiters hunting engineers where LinkedIn alone falls short — GitHub, Stack Overflow, forums, conference archives
- Enterprise HR operations teams that require SSO, SCIM, role-based access, audit logs and EU data residency
When not to use Findr
- Buyers who need a published price list: nothing is quoted anywhere, every path leads to a sales conversation
- Anyone wanting to self-serve or trial the product alone — no free plan and no free trial is advertised
- Legal and procurement teams that require published contractual documents before evaluating: no privacy policy, terms or DPA is available on the site
- Teams looking to automate interviews, offers or closing — Findr deliberately stops at the handoff to humans
- Developers expecting a public API or a mobile app for the business product; neither exists
How to use Findr
A typical end-to-end flow, from setup to results.
- Request a live demo through the booking link, or contact the sales team by form or email
- Take a 30-minute session in which Findr runs a real search on one of your open roles
- Review the shortlist produced during that session and the tailored playbook shared afterwards
- Agree commercial terms with the sales team, since no price is published
- On day one of onboarding, connect your ATS, calendar and inbox
- Set up SSO, SCIM and role-based access if your organisation requires them
- Define the rubric and the playbook for your first role, including approval and exclusion rules
- On day two, brief Marvin on that role by pasting a job description or describing it in plain language
- Answer the agent's clarifying questions, then review the first shortlist, usually within a day
- Approve outreach before it goes out, then collect booked interviews and one-page briefs from the dashboard
Pros & Cons
Pros
- Covers the whole top of the funnel end to end rather than only the search step
- Sources well beyond LinkedIn, which matters for hard-to-find technical profiles
- Scoring is explainable: the reasoning is shown, not just a match percentage
- Human approval before outreach is the default, and the per-role playbook is inspectable
- Bias prevention is described concretely, down to stripping identity before any model reads a profile
- EU data residency by default on Google Cloud Frankfurt, with encryption in transit and at rest
- Broad, explicitly named integrations across ATS, email, calendar, chat, SSO and data warehouses
Cons
- No price is published anywhere and there is no pricing page, so the cost cannot be assessed before contacting sales
- No privacy policy, terms, cookie policy or DPA is published; the footer links carrying those labels are inert text
- SOC 2 Type I and Type II and ISO 27001 are claimed as active without a certificate number, a named auditor or any report
- Headline performance figures are given with no methodology, scope or date
- No postal address is published, and the company name shown in the footer cannot be found in the Spanish commercial register
- No free trial and no free plan is offered, so the product cannot be evaluated without a sales process
- No public API and no mobile application for the business product
Pricing & Plans
Findr publishes no pricing whatsoever. There is no pricing page, no entry for pricing in the navigation, and no amount appears anywhere on the site; the only euro figures shown are illustrative candidate salaries inside product mock-ups, not the cost of the tool. The site describes its commercial approach only in qualitative terms: usage-based pricing for mid-market and smaller companies, and billing by role, by month or by seat for agencies and RPO providers, while enterprise buyers are told the company is procurement-ready with a DPA, an MSA, SOC 2 and ISO 27001 documentation. Neither a permanent free plan nor a free trial is advertised. Every call to action leads to booking a demo or talking to sales, so the entry price must be obtained directly from the vendor.
Data, GDPR & hosting
A consolidated view of how Findr handles your data.
GDPR overview
Findr makes explicit GDPR claims but publishes nothing to support them. The Security page lists GDPR compliant as an active status, describes privacy by design, offers a DPA on request, and states EU data residency on Google Cloud in Frankfurt by default, with TLS 1.3 in transit and AES-256 at rest. It also describes stripping identifying fields — name, email, phone, gender, date of birth, age and photo — before any model reads a candidate profile. Against that, no privacy policy, cookie policy, terms or DPA is actually published: the footer links carrying those labels are inert text. No data protection officer and no Article 27 representative is named, and no postal address appears anywhere. The publisher is established in Spain, so it falls directly under the GDPR.
Who owns the data?
No terms and conditions or privacy policy is published, so ownership is not defined contractually anywhere a reader can consult. The only statements available are marketing claims on the Security page, which asserts that your data stays yours, that each customer is isolated in its own tenant with no shared model state, and that data is deleted end to end within thirty days of a request. Candidate and pipeline data is said never to be used to train third-party models. Because the footer links to Privacy, Cookies, Terms and DPA are inert placeholders rather than real pages, none of these commitments is currently backed by a document a customer could rely on or enforce.
Reuse rights
Nothing on the site grants or restricts a customer's right to reuse the data the platform produces, because no terms of service exist to define it. In practice Findr describes an outbound flow rather than a dataset the customer takes away: Marvin sources candidates, scores them against a rubric the customer defines, writes outreach in the customer's own brand voice, and writes results back into the customer's own ATS, calendar and Slack. Full reasoning traces, scores and audit logs are said to be exportable, including to a SIEM, which implies the customer can retrieve what the agent produced. Findr states that candidate and pipeline data never feeds third-party model training, but says nothing about its own models, and publishes no document confirming any of it.
Data retention & training
Hosting summary
Findr states that EU data residency is the default, with hosting on Google Cloud in the Frankfurt region, and presents this as a standing arrangement rather than a paid option. Data is described as encrypted in transit with TLS 1.3 and at rest with AES-256, and each customer is said to be isolated in its own tenant with no shared model state between customers. The customer-facing application additionally relies on Google Firebase, which is visible in the public configuration of the Findr dashboard. Beyond that, detail is thin: no list of subprocessors is published, although the company says one is available on request alongside a DPA, an MSA and a security questionnaire. No privacy policy or contractual document sets out the hosting arrangements, so all of the above rests on statements made on the Security page rather than on anything a customer could enforce.
Things to keep in mind
Risks and trade-offs to weigh before adopting Findr.
- Candidate data is among the most sensitive personal data an employer handles, yet no privacy policy, terms or data processing agreement is published to govern how Findr processes it
- Bias claims are strong but unverifiable from outside: stripping identity and running several agents in parallel reduce risk without eliminating it, and an automated rubric can still encode the preferences of whoever wrote it
- Delegating first contact and screening can quietly erode a team's own sourcing judgement, leaving recruiters dependent on an agent whose reasoning they may stop reading
- Security certifications are asserted without certificate numbers, a named auditor or any report, so they cannot be relied on in a risk assessment as they stand
- Candidates are screened and scored by an agent that may be perceived as opaque; automated decisions about people carry legal and reputational exposure that the site does not address
- Because no price is published and no trial exists, cost and commitment can only be discovered inside a sales process, which makes budgeting and comparison harder
- The publisher's identity is inconsistent: the name shown in the footer cannot be found in the Spanish commercial register, which complicates due diligence and recourse
Setup & Integrations
Technical difficulty
Low for the customer's technical staff. Findr describes onboarding as two short sessions: connecting the ATS, calendar and inbox on the first day, then briefing the agent on a first role on the second. There is no installation, no migration and no parallel system of record, since the agent writes into tools the team already uses. Larger organisations add single sign-on and SCIM provisioning, which requires IT involvement but follows standard patterns. The real effort is not technical: it lies in defining the scoring rubric and the per-role playbook well enough for the agent to act on them.
Deployment
Integrations
Supported languages
Behind Findr
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Findr.
Frequently asked questions
What exactly does Findr's agent Marvin do?
How is this different from a sourcing tool such as LinkedIn Recruiter?
Does it replace recruiters?
How much does it cost?
Where is the data hosted?
Is candidate data used to train AI models?
Does Findr publish a privacy policy, terms or a DPA?
Is there an API or a mobile app?
What does it integrate with?
How long does it take to get started?
Should you pick Findr?
Findr addresses a genuinely painful and well-chosen problem. Top-of-funnel recruiting is repetitive, high-volume work that burns out sourcers, and an agent that reads public signal, reasons about fit, writes in your voice and never stops following up is a credible answer to it. The scope is sensibly drawn: Marvin stops at the handoff, leaving interviews, offers and closing to humans. The functional depth is real, the integration list is specific rather than vague, and the bias-prevention section is more concrete than most — stripping identifying fields before any model reads a profile, and running models deterministically, are architectural commitments rather than slogans. The product is demonstrably in service: a working application runs on a separate dashboard with account creation. The reservations are about evidence rather than ambition. SOC 2 Type I and Type II and ISO 27001 are displayed as active certifications with no certificate number, no named auditor and no report. Performance figures are quoted without methodology or scope. Customer logos appear without a single case study, and the testimonials are anonymous. Most consequentially for a tool handling candidate data, no privacy policy, no terms, no cookie policy and no data processing agreement is published at all — the footer links carrying those labels are inert text, and an archived version of the privacy page was filled with placeholder Lorem ipsum. The company name in the footer, Findr Labs, S.L., does not appear in the Spanish commercial register. Findr is worth a conversation for a talent team with real volume, particularly in technical hiring. Go in expecting to obtain pricing, contractual documents and certification evidence directly from the vendor, because the website supplies none of them.
- Choosing a selection results in a full page refresh.
- Opens in a new window.