What's new in Goodfit

Letting an interviewer see a candidate used to mean making them a recruiter on the job, with every phone number, form answer and team note that comes with it. Now you add them to the job's interview panel instead.

• Add people under Team → Interview Panel on the job's edit page. Anyone who isn't on your team yet gets an email invite.
• They open a candidate's report and submit a scorecard. They can't move, reject, hire, email, share or download a candidate.
• Choose what each person sees. Five switches: Contact details, Application form answers, Resume, Team notes and Other interviewers' scorecards. Start from Panelist (resume only) or Hiring manager (everything except team notes), then adjust.
• Booking an interview adds them for you. When you schedule, everyone invited who can't open the report yet is listed and ticked. Add and schedule adds them to the panel, so the report link in their calendar invite opens. Schedule without adding books it as before.
• No peeking at colleagues. An interviewer sees the other scorecards only once they've submitted their own, so nobody anchors on someone else's verdict.

A new kind of cheating tool reads the question off the candidate's screen and types the answer in for them. There's no paste, no tab switch and nothing in the browser for our other checks to find. We now catch it by how the code arrives.

• Code that appears far faster than anyone types, with the mouse not moving, is flagged in the Integrity Report as Automated typing detected, with the time so you can watch that moment in the recording.
• It's a high-risk flag. The attempt's Integrity card leads with it, in red.
• Fast typists are safe. The bar is about twice the speed of a quick typist.
• Candidates are told, not stopped. The bar at the top of the assessment reads Unusual typing pattern detected. Recorded. We let them carry on on purpose: the more of the session we see, the clearer the picture you get.

General questions get rehearsed answers full of the right keywords. Now the interviewer can ask each candidate about the work they say they did.

• Switch it on per interview. In the AI interview builder, turn on Resume Based Questions in the Resume row, between Screening and Competencies.
• Up to 3 questions, about real achievements. The interviewer agent picks specific things from the resume, like a sales record, a team they led or a named project, and asks what the candidate did themselves and what came of it. A thin resume gets fewer questions, and no resume means none.
• Never part of the score. Each answer is rated Strong, Average or Weak on its own. It doesn't count toward the interview score.
• See it on the candidate report. A Resume Questions card puts what the resume says next to how the candidate answered, with a jump to that moment in the recording. When the two don't line up, you'll see Doesn't match resume with the candidate's own words.

A dead mic, a dropped connection, a closed tab. A candidate who joined and never said a word used to get a score of 0 and no way back in. Now they get another go.

• Not scored, not charged. If the candidate doesn't answer a single question, the call gets no score, uses no credit and doesn't move them through the pipeline.
• The candidate gets an email to retake it. Same interview, same link. The email says the call didn't go through and that nothing counts against them. At most one a day, so a faulty device never floods their inbox.
• Marked Incomplete on the candidate's page. The interview card reads Incomplete and says New link sent., or No new link sent. when the email couldn't go out.
• Watch what happened. Click the card for the reason when it's clear (a weak connection, the candidate leaving, or the tab being closed) and a video of each failed call. Once the retake is scored, the earlier calls stay one click away under Incomplete calls.

One total score doesn't tell you whether someone is strong at Excel and weak at customer communication. And when hundreds of people sit the same test, the questions get shared.

In skill assessments

• Put questions into skills. In the question list, open the menu and pick Add skill head. Name it (say, Excel) and add questions inside it, straight from Add question or the Question bank.
• Ask a few from each skill. Click Asks all 5 · change at the bottom of a skill and set how many to ask. Every candidate gets a different mix, so answers are harder to pass around.
• The report scores each skill separately. A Skills section at the top of the report shows points and a percentage per skill, then the questions under each one.

In AI interviews

• Set it per competency. In the interview builder, click Asks all 5 · change under a competency and pick how many to ask.

Hiring someone for a Systems Engineer role? Test them on the kind of work they actually do: reading logs, pulling out columns, counting and sorting text.

• New Bash questions in the bank. 50 of them, ready to drop into an assessment.

Describe a rule in a sentence and the AI fills in the trigger, the conditions and the actions. When your sentence leaves something open, it now asks you instead of picking for you.

• One question at a time, with the answers ready to click. Say "reject anyone whose AI interview score is too low" and it asks where the line is, offering 3, 4 and 5. Pick one and the rule updates in place.
• Each question points at the card it will change. Trigger, Conditions or Actions lights up while you answer, so you can see where your answer lands.
• It fills its own gaps too. If the rule needs a stage, an email template or a number of days and your sentence didn't say, it offers the real ones from your pipeline instead of leaving you an empty dropdown. Other takes any number you want.
• Skip and Skip all are always there. The questions are a shortcut, never a gate, and everything stays editable by hand below.

Rules that reject a candidate ask you first

• Anything that rejects, or marks someone rejected, shows Confirm before publishing and says plainly: This rule rejects candidates without a human looking first.
• Two safer answers are one click away. Flag my team instead swaps the rejection for an alert to the recruiters on that job. Mark as poor fit instead sorts the candidate on the board without closing them out. Keeping the rejection is something you choose, not something you drift into.

It tells you when a rule cannot do something

• Ask for something automations don't do yet, such as deleting or archiving records, and you get That isn't something a rule can do yet with the reason. Nothing is filled in below, so you can describe a different rule or build one by hand.

Don't want to open the dashboard to find out how someone did? Put Goodfit in a Slack channel and the results come to you.

• Connect from Settings → Integrations → Slack. Pick your workspace, pick the channel, hit Allow. No keys to copy.
• Every result posts as a card — the candidate's name and score, which assessment and which job, their email, how long they took and the language they spoke, and an Open Report button straight to the full report.
• You choose what gets posted. Interview results covers every interview and assessment type and is on from the start. New applications posts people who applied on their own or came in through a referral, and is off until you turn it on. You don't get notified when the candidates you invited start the assessment.
• Send test message drops a sample card in the channel so you can watch it land. Change channel moves the feed, Disconnect ends it.

Someone picks English, gets two questions in, and realises they'd do better in Hindi. The interviewer can now switch languages if Allow multi-lingual responses is enabled.

• Candidates can request a language change at any time. For example, if someone says, "Hindi mein kar sakte hain?" the interviewer will seamlessly continue in Hindi from the same question, provided Hindi is enabled as a supported language. The interview carries on without interruption.

• You control whether this is possible. Turn on Allow multi-lingual responses in the interview builder to enable this feature. If left off, the interview will stay in the original language selected by the candidate.

A candidate who switched is notified immediately

• A small bar drops in at the top of the assessment carrying the running count: You switched away. Recorded, 3rd time.
• Coding, skill-based and language assessments hold until they're back in fullscreen but, for AI interviews the bot won't pause even if fullscreen is disabled

The coding question bank just doubled: from 525 to 1,000 questions, and the new ones bring more than just quantity.

• 1,000 coding questions, ready to use across screens, algorithms, SQL, puzzles, at easy, medium and hard.
• 427 questions now show you, not just tell you, what they’re about. All SQL questions include visual data cards — see what’s in each table, field types and keys, not just paragraphs of text. Grids, family trees, timelines, step-by-step chains: they're drawn out, examples shaded, so candidates can see the problem without having to picture it.

• All five tests — Big Five, MBTI, DISC, work values and emotional intelligence — read end to end in Hindi, Marathi, Telugu, Kannada, Malayalam, Bengali, Punjabi, Tamil, Odia, Gujarati and Assamese.
• Scores don't move. The same answers in Marathi and in English produce the same result.

Hiring on Keka? Link a Keka job to a Goodfit assessment and the loop runs itself: candidates flow in, get screened, get invited, and results land back on their Keka profile.

• Connect in two minutes from Settings → Integrations with API credentials generated in your Keka account. Nothing to install on the Keka side, and demo tenants work too.
• Import the Keka job in one click. Title, description, locations and seniority arrive as a draft job. Add your hiring stages and assessments, publish, and syncing starts.
• Every synced candidate lands in Prospects with their Keka resume already pulled and screened: a score out of 10 and the reasoning, before anyone has sent an email.
• Invites fire on your trigger: the moment a candidate is added to the Keka job, when they reach a pipeline stage you pick, or by hand with Invite now.
• Results return to Keka on their own. The resume screening score arrives as a note the moment it's ready; the final result follows with section scores, proctoring flags and a link to the full candidate report.

Every call is recorded, and the whole interview sits on one card.

• Every call has its own recording and its own player.
• One card per phone round. One call or five, it's the same card and the same score.
• The transcript reads as one conversation. Where a call dropped, a line in the transcript says so, and says how long it was before we rang back.
• 5 of 6 questions asked sits next to the score whenever the interview ended short.

A dropped call used to mean starting again from the greeting: the same questions in the wrong language, a second half-finished interview, and two scores where there should have been one.

• The call back picks up where the call stopped. Questions already answered, or declined, are not asked again, and neither is the ice-breaker.
• It opens by saying why: the earlier call got cut off, is this still you, is now alright. In the language they'd already chosen.
• One score for the whole interview, worked out from both calls together. One interview, charged once.
• If the call back never connects, we score the answers we did get.

A phone interview used to have to be the entire pipeline. Now it can be the first round of one.

• Any stages you like can follow the phone interview. It stays first.
• Auto Progress, on the phone stage, decides what happens when the call ends. On: the candidate moves to the next stage and is sent the link. Off: their card waits in the phone column until you move it.
• The link goes out by email, and by WhatsApp as well when the job sends WhatsApp invites and the candidate has a usable number. The email names the round they're being sent to.

If multiple languages are enabled for an assessment, the bot asks the candidate which of those languages they'd like to continue in, before the interview starts. During the call, the candidate can switch to any of the enabled languages at any point.

• Ring it back and, if there's a call waiting to be made to that number, the interview starts there and then, without waiting for the next attempt.
• If their earlier call dropped, ringing back continues it from where it stopped, exactly as if we'd rung them.
• The number is 080 3180 3101. Ring it with nothing scheduled and there's one short line and the call ends.

Four percentages and a paragraph tell you what the candidate scored. They don't tell you what it's like to work with them. Every psychometric report now carries a written profile underneath the numbers.

• Behaviour at work — how they decide, communicate, collaborate and lead, in plain sentences rather than trait names.
• Under pressure — where they hold up, where they slip, and what a manager should do about it.
• Trait by trait — what each score actually means, the strengths that come with it, and the watchouts.
• Motivators, blind spots and ideal work environment — what they'll work best with, and what they're less suited for.
• Role families they suit, and six interview questions worth asking next.
• Same depth for every framework: Big Five, MBTI, DISC, work values and EQ all read the same way.

• The score strip now leads with a real fact instead of a status word: the DISC blend and its name, the MBTI archetype, the EQ score, the strongest Big Five trait, the top work value.

And three things that give the numbers an anchor they never had:

• Fit for this role. The job description sets a target band per trait, once per role, and every applicant to that role is measured against the same bands. Each trait reads in range, above range or below range, with how many points off, plus how many traits land inside and which one sits furthest away.
• Cohort percentiles. 91st percentile · vs 238 candidates for this role. Once a job has enough evaluated candidates the comparison is against that job's own pool; below that it's against everyone who has taken the assessment. The pool and its size are printed next to the number, so you always know what the rank is a rank of.
• Response quality. A profile only means something if the candidate engaged with the questions. Each report carries a band — Good, Read with care or Low confidence — worked out from pace, how much the answers varied, how often they sat on the middle option, and whether they answered consistently. It says which of those triggered it.

• On Low confidence the percentile is withheld and the profile carries a caveat, rather than quietly presenting a rank built on answers nobody trusts.
• Everything above is in the downloadable candidate report too, not only on screen. Reports evaluated before this update render exactly as they did.

Scheduling used to be a guess: pick a day, pick a time, invite three people, hope none of them are double-booked. The schedule dialog now shows you.

• Pick a day and every invited teammate gets their own lane. Busy time is drawn as blocks with their name on it; free time is simply empty.
• Everyone free slots appear as chips under the time inputs. Click one and the start and end times fill in, at whatever duration you'd already set.
• Pick a time that clashes and a line tells you who it clashes with. It's a warning, never a block, because sometimes you know something the calendar doesn't.
• Teammates who haven't connected a calendar show as availability unknown. Nobody is blocked from being invited.
• Reads the calendars your team has already connected, Google or Microsoft. Nobody reconnects anything and nobody grants a new permission.

The AI interviewer speaks beyond the Indian language set now.

• Japanese and Arabic are now available, expanding the list of supported languages alongside English, Hindi, Marathi, Telugu, Kannada, Malayalam, Bengali, Gujarati, Punjabi, Tamil, Odia, and Assamese.

The lobby checked the camera, the speakers and the microphone, and then let someone on a dying connection start anyway.

• Network joins Permissions, Speaker, Microphone and Recording in the pre-interview checks. It runs quietly in the background while the candidate does the others, so the step is already settled by the time they reach it.
• A weak connection warns, shows what was measured in terms a candidate can act on (Speed and Delay), lists four things to try, and still leaves Continue anyway available.
• A connection too weak to carry a call at all doesn't get to start one.

The funnel could tell you how many candidates reached a stage. It couldn't tell you how they did there.

• Candidate Flow is a new view on every job, reached from the board. It draws your Kanban as a flow: where candidates came in, which stages they passed through, and where they're sitting now. Every column reconciles with the board it came from.

• The org-wide funnel splits its old undifferentiated "Active" band into Goodfit and Poorfit, so "how many of the people who got this far were actually any good" finally has an answer.
• A Stage Outcomes table sits under it: per stage, how many entered, how many came out good fit or poor fit, and how many were hired, rejected, withdrawn or are still in play.
• Team Activity gains jobs opened, jobs closed and closed with a hire for the period you're looking at, and Avg Decision (days) now shows a number instead of a blank.

Human interviews were the one part of the pipeline with no numbers on it at all. Insights → Interviews is a new page.

• How many interviews were booked over any date range, how they ended, and how often they were rescheduled.
• How much notice interviews are typically booked with, and where they came from.
• Broken down by job and by the person who scheduled them.

Agencies source for you, and until now the only role that fit gave them everything: your own candidates, and every rival agency's candidates, on every job they touched.

• Recruitment Vendor is a new role. It sees every open role in the organisation, and only the candidates their own agency submitted. Nothing else on the board exists for them.
• The agency is a real thing you name, so five people from one firm roll up to one agency. Set it while inviting them, or from the members list afterwards.
• Vendors submit candidates and follow their status. They don't move cards between stages, reject, hire, or email candidates, because the party being measured shouldn't be the party keeping score.
• Your side gets a Sourced by filter on the board and a chip on the candidate card, so you can always see which agency sent whom, and job cards count only what the viewer can see.
• Vendors can't reach members, billing, email setup or API keys, and can't edit a job.

The same person could end up on the same job twice: two cards, two invitation emails, two sets of results, and a recruiter working out which one was real.

• A candidate who is already live on a job is no longer added a second time, whichever door you come in through: resume upload, bulk invite, a board re-invite, or a single manual invite.
• Re-inviting someone who is already moving through your pipeline now tells you so, and points you at their existing card rather than silently starting a parallel application.
• Bulk invites report what they skipped. The count is real now, where before the toast could only ever say zero.
• Someone you rejected or closed out earlier is a legitimate fresh application. Invite them individually and you're asked to confirm first, with the button reading Invite Again. A bulk list skips them rather than guessing on your behalf.

The jobs list could be narrowed three ways and remembered none of them. Leave the page and it was gone; refresh and it was gone.

• The jobs list now uses the same filter builder as the board and the candidates page, across twelve fields, including seniority, workplace type, locations, skills and tags. "Senior remote roles" is now something you can express.
• Save a set of filters as a named view and it becomes a tab. Views carry their filters, their sort and their columns, so a view can show different columns as well as fewer rows.
• Rename, duplicate or delete a view from the ⋯ menu beside the tabs, which is also where you save whatever you're currently looking at.

• Come back to the jobs list later and your last view is waiting, per organisation. A link with a view in it always wins.
• Views never save themselves behind your back. Change one and you get Save, Save as new and Reset; the view's own filters sit dimmed so you can see what you added on top.
• A Columns picker decides what the table shows, and reorders it. Six columns are on by default. Views are private to you.

Interviewing well and writing everything down at the same time is a trick nobody actually pulls off. One of them always gives. The Meeting Assistant joins your video interview, records it, transcribes it, and has a write-up waiting on the candidate's profile by the time you're out of the call.

• Works with Google Meet and Microsoft Teams. Nothing to install, and you don't have to connect your calendar for it to work.
• Schedule an interview as you normally would and the assistant is already on the invite — you'll see Goodfit Interview Assistant sitting in the participant list next to your interviewers. It joins at the start time like anyone else.
• Recording is on by default. To skip it for one interview, switch Record & AI evaluation off or take the assistant out of the participant list — either one, they're the same switch.
• Forgot to include it, or the meeting was set up outside Goodfit? Invite assistant@apply.goodfit.so to the calendar event exactly as you'd invite a colleague, and it will turn up. There's also Add meeting link to record on the candidate's card if you'd rather paste the Meet or Teams link.
• The interview card tells you where things stand without you opening anything: Will record, Recording, Recorded, and — for the cases worth knowing about — Not admitted when nobody let it into the call, or Recording blocked when the meeting's own settings wouldn't allow it.

Open the candidate and the interview card holds everything the conversation produced.

• The recording, playable in the browser or downloadable.
• The full transcript, with speakers separated. Click any timestamp and the video jumps to that moment.
• An AI write-up: what was covered, how the candidate did against each competency with the reasoning behind every score, strengths, concerns, and anything worth a second look. Every claim carries the timestamp it came from — click it and the recording scrolls into view and plays from exactly there, so you can check the evidence yourself in about two seconds.
• Bilingual interviews transcribe properly now, rather than forcing the whole recording into one language.

Feedback that arrives as "yeah, decent" in three different chat threads isn't feedback you can compare. Now everyone answers the same question.

• After the call, everyone marked as an Interviewer gets an email with a private link: Strong No · No · Neutral · Yes · Strong Yes, plus per-skill ratings and notes. No account, no login — external panellists can use it too.
• Mark someone as an Observer while scheduling and they're left alone. Only the people who actually interviewed get asked.
• Every verdict lands on the candidate's profile next to the recording, so you can see where the panel agreed and where it didn't.
• Candidates get a short optional survey about how the interview felt — whether the interviewer was prepared, whether they got to show their skills, whether the role was explained. It has no bearing on their application, and they never see your notes, the recording or the write-up.

A second interview used to quietly replace the first one. Two rounds meant losing the record of the round you'd already done.

• Schedule as many video calls with a candidate as you like. Each one gets its own card, its own recording, its own transcript, its own write-up and its own panel feedback — the earlier rounds stay exactly where they were.
• Moving a candidate between stages no longer cancels a booked interview or its recording. That includes rejecting or hiring them, which used to pull the interview out from under a call that was still going ahead. If a meeting shouldn't happen, cancel the meeting.
• Reschedule and the assistant follows the new time. Cancel the event and the recording is cancelled with it. Change the attendees and the update carries through.

Recruiting lives across half a dozen tabs, and most of the day is spent copy-pasting between them. The new Goodfit browser extension sits in those tabs with you: sign in once and your Goodfit jobs and candidates are right there on LinkedIn, Indeed and Wellfound.

• Sign in with your Goodfit account once — the same connection works across every supported board.
• Everything is started by you and only touches your own organization's data. No ads, no browsing tracking, nothing sold.

Open a board's "post a job" page and the form is already filled in from your Goodfit job.

• Pick a job in the extension, choose the board (LinkedIn, Indeed or Wellfound), and the posting form opens pre-filled — title, location, pay, job type and description.
• Review each step and submit it yourself. The extension never submits, sponsors or pays for a post on your behalf — you stay in control of publishing and budget.

On a board's applicants page, bring the candidates into a Goodfit job in one flow.

• Click Import applications to Goodfit, pick the organization and the job to import into (only active jobs are listed), choose a batch size, and go. Résumés are processed in the background.
• Rejected applicants are never imported, and imported candidates are tagged with the board they came from so you always know the source.
• Prefer to move faster than prospecting? Tick "Invite candidates (send Goodfit assessment) instead of adding as prospects" to send the assessment invite as you import.

Already have jobs live on a board? Import them instead of recreating them by hand.

• From the board's jobs list, tick the ones you want (there's a search box for long lists) and import them into Goodfit as drafts, ready for you to review before activating.
• Jobs you've already imported are skipped, so re-running is safe.

Sourcing on LinkedIn? You don't need to leave the profile.

• On any LinkedIn profile, the extension shows whether that person is already a candidate in your account, so you don't re-add or re-contact the same person.
• Add them to a job as a prospect, or send them a Goodfit assessment invite — right from the profile.

Booking an interview used to mean writing the time into Goodfit, then writing it again into your calendar, then sending the invite yourself. Connect your calendar once and Goodfit does all three.

• Connect under Settings → Personal → Connected Apps. From then on every interview you schedule from the applications board becomes a real event on your own calendar — Meet link, candidate invite, reminders, the lot.
• You are the organiser, so the candidate sees your name in the invite and their reply lands in your inbox — not in a robot mailbox nobody reads.
• Reschedule or cancel from either side. Drag the event in your calendar and Goodfit keeps up; change it in Goodfit and your calendar does.
• Don't want to connect? Schedule without connecting still books the interview in Goodfit with its time, its attendees and its notes. Nothing about scheduling is held hostage to the integration.
• The week reads back on your dashboard: a Meetings card with the next 7 days, grouped by day, your team's interviews and not only your own, each with a one-tap join button. Nothing scheduled, no card.

A four-stage pipeline tells you what each stage is. It never told you what the whole thing does — who gets filtered, at what score, and where they end up. Now there's a summary that says it in sentences.

• A How candidates move strip sits under your stages: one tile per step, in order, so the journey reads left to right. Open it for the step-by-step version.
• Every line names the consequence — "Scoring under 12 out of 30 lands them in your poorfits. Everyone else moves on." — and closes with where a candidate who finishes everything ends up.
• Steps with a problem are ringed and counted in the headline. The one this was built for: a cut-off switched on but left at 0. The switch reads "On" and filters nobody, and until now nothing said so.
• It updates as you type, so you can check your own understanding without publishing and watching real candidates.

A timer that looks fine next to a question count can still be far too short for the questions themselves — and the candidate is the one who finds out. Goodfit now does that arithmetic for you, before you save.

• Save or publish and anything that doesn't add up gets one dialog with the arithmetic spelled out — Candidates get 1 minute, but the questions add up to about 40 minutes — that's about 30 seconds per question.
• Timer far shorter than the questions need? Use suggested time (40 min) applies the right number and saves in the same click.
• Per-question times are judged by what you're asking for: 15 seconds to pick an option is fine, 15 seconds to record a video answer is not.
• A pass score of 100 (only perfection passes) or 0 (everyone passes) gets flagged too, in case it wasn't deliberate.
• Save anyway is always there. Only settings no candidate could ever take — a 0-minute timer, zero questions — are blocked outright, because there's nothing to confirm.

If your organisation sends candidate email from your own address and it stops working, candidates quietly stop receiving invitations. The person who notices is a recruiter. The person who can fix it holds the mail-server password. Now there's a route between them.

• Settings → Email shows a status card when sending is failing — and nothing at all when it's healthy, so the day it appears, it means something.
• The card names what actually broke and what to do about it. A wrong password and a server refusing to send now get different advice, instead of the same list of steps half of which don't apply.
• Send to IT emails your mail administrator the whole picture: the server, the username, the sending address, the exact error, and numbered steps to fix it.
• Pick a teammate from the list, or type any address — someone outside your team gets invited, and the invitation itself explains the outage rather than arriving as a bare "you've been invited".

The 20 credits every new organisation starts with used to sit there forever. They now have a deadline, and you can see it counting down.

• The credit meter in your sidebar shows the days left alongside your balance — amber at three days, red on the last one.
• Two emails, no more: one warning three days out, one the day the unused credits go.
• Only what you haven't spent expires. Anything you've already used is untouched.
• Buy credits at any point and the clock stops for good — purchased credits never expire, and the countdown doesn't come back.

Everything we watch during an assessment — the camera, the browser, the answers, the voice — used to live in two separate panels with their own language for "suspicious". Now every signal lands in a single report with a single verdict.

• Each assessment attempt carries one Integrity card in its header — HARPA AI extension, AI-written answer (95%) + 35 more flags. Open it for the full report.
• The report opens with the verdict in words, not a score to decode — and when a face doesn't match, the two photos sit side by side at the top so you can judge it yourself in a second.
• A timeline of the whole session shows every flag on the moment it happened, labelled — Earbuds, Looking away ×2 — and for voice interviews, tap any moment to jump the recording straight to it.
• A What we checked grid lists all eight checks for that assessment type, so you see what passed as clearly as what didn't. Nothing wrong? You get "All 8 checks passed" instead of a wall of empty tables.
• Below that, everything with its evidence attached: AI-written answers with the exact quoted passage, voice-to-lip match, the flagged camera frames, and a tidy event log you can filter to browser, camera, microphone, or tab.

A candidate can install Monica, HARPA, Sider or the like and have an AI assistant answer questions in a floating panel — same tab, no tab switch, nothing for the old checks to see. We now catch it.

• If a candidate had a known AI assistant open during their assessment, it turns up in the Integrity Report by name and logo — nothing for you to switch on, and nothing the candidate is told.
• We separate used from merely installed: an assistant that was open and answering is a red, high-risk flag; one that was installed and never opened is an amber "installed, no visible use" — worth knowing, not an accusation.
• Every flag comes with the evidence behind it — how long the assistant was on screen, and what it was showing while the candidate worked.
• Covers all four assessment types — coding, skill-based, psychometric, and voice interviews.

Identical question sets leak — one candidate finishes, the questions land in a WhatsApp group, and the next twenty walk in prepared. Now you can attach a bigger set and hand each candidate a random slice of it.

• Split your questions into Fixed — everyone gets these — and a Pool everyone draws from. Click the pill on any question card to move it between the two.
• Set how many to draw from each difficulty: 1 easy, 2 medium, 1 hard. Every candidate gets the same shape of test, just not the same questions.
• A summary bar spells it out as you build: "Each candidate receives 6 questions: 2 fixed + 1 easy + 2 medium + 1 hard from the pool". No mental arithmetic.
• Attach up to 24 questions, deliver up to 6 — and reshape the pool right up until a candidate starts; your changes count for everyone yet to begin.
• Scoring is unchanged — the cutoff is still "solve X of the questions delivered", and it now counts against what the candidate actually got, not everything you attached.

Someone dropped the cutoff on the coding round, or a stage vanished from the pipeline, and nobody could say who or when. Every job now has an Activity page that reads like a sentence.

• Open it from the clock icon on the job's applications board — a plain-English timeline of every change, newest first.
• Pipeline changes read like people talk: "Lowered cutoff 2→1 on Clawbot Coding Assessment", "Added 'Web Interview' as Stage 6", "Moved 'CEFR English' to Stage 3". Open any row for the stage-by-stage detail.
• Job changes are there too — published, closed, made public or private, details edited, duplicated.
• Every row names the person who did it and the exact time, alongside a friendly "10 min ago".
• Filter to Pipeline or Details & visibility when you're chasing one kind of change.

Our photo check proves the candidate's face and name match the ID they're holding up — but not that the ID itself is real. DigiLocker closes that gap: the photo and name come from UIDAI, so the document can't be forged.

• On the ID step, the candidate picks Connect DigiLocker (recommended) or Take a photo of your ID instead — the fallback stays for anyone who doesn't use it.

• They give consent in DigiLocker, and we receive their Aadhaar photo and name from the government source. We match the photo against the selfie they just took and the name against the applicant on file.

• A pass sends them straight into the interview. A mismatch is flagged for you to review — same as today.

• You see it in the same verification panel you already use: a via DigiLocker badge and the photo labelled as pulled from DigiLocker.

Voice interviews already watch the camera during the call, but a quick camera-flip to a friend, a phone just out of shot, or someone feeding answers can slip through. Now every voice interview gets a careful second look once it's over.

• Every voice interview is reviewed automatically after it ends — nothing to switch on, nothing for the candidate to do, and zero effect on their call.
• It looks for the things that actually matter: a different person answering, an extra person in frame, a phone or earbuds, reading off notes, or a photo held up to fool the camera.
• A calm Video Proctoring summary lands on the candidate's voice assessment — a single reassurance line when nothing's off, and a clear verdict with the flagged moments when something is.
• Open the full report for a timeline of the whole interview, what we checked, and the exact flagged frames — tap any moment to jump straight to it.

Closing a filled role used to just flip a status — and the hire never made it into your numbers. Now closing asks the one question worth asking, and pre-fills the answer.

• Hit Closed on a job and a short Who did you hire? dialog opens, pre-checked with anyone already marked hired or holding an accepted offer — best case, you confirm and you're done.
• Hired someone from the pipeline? They're recorded as a hire, so time-to-hire, conversion, velocity, and sourcing stats fill in for free.
• Hired someone off-platform? Add their name — and email if you have it — as many people as you need.
• Didn't hire anyone? Pick a reason — filled internally, budget freeze, cancelled, still deciding — and leave a note.

A coding assessment built for a specific job used to open as a blank bank you searched by hand. Now it opens with three questions already matched to that role.

• Create a coding assessment for a job and the empty state leads with 3 suggested questions, ranked against the job's title, description, and required skills.
• One Add 3 suggested tap drops all three in — then add, swap, or edit like any other question.
• Picks come straight from your existing question bank, so they're questions you already trust — surfaced up front instead of hunted for.
• Building a standalone test with no job attached? The plain builder is exactly as it was.

Every Insights page already exported a CSV — but if you live in spreadsheets, that meant download the file, open Sheets, import it. Now it's one click.

• A new Open in Sheets button sits next to Download CSV on every Insights page — overview, pipeline, activity, and sourcing.
• It opens the data you're looking at as a live Google Sheet — one tab per section, with bold, frozen headers.
• It matches your filters and date range exactly — what's on screen is what lands in the sheet, nothing more.
• The sheet opens shareable by link, so you can hand the numbers to anyone who needs them.

• A new Save For Later button sits beside Start Interview on public job pages.
• One tap opens a Google Calendar reminder, pre-filled and linking straight back to the role — the candidate picks whatever time works for them.
• It's a personal reminder, not a booking — nothing is reserved and no one on your team is pinged.

A voice scorecard told you what a candidate answered. Now it can also tell you how well they speak the language the interview ran in.

• Flip on Score language proficiency in the voice assessment's settings.
• Broken out across grammar, vocabulary, fluency, and coherence, each with a reason and a short summary.
• It's judged in the language the candidate actually spoke, answer fluently in Hindi and you're scored on Hindi, no English penalty.
• The proficiency folds into the candidate's overall /10, so it counts toward the same cutoffs as every other skill.

The person who first posted a job often isn't the one who should be fielding its activity a month later. Now you can hand it over.

• A new Ownership card on the job's edit page shows who currently receives the job's notifications, with a Transfer button.
• Hit Transfer, search your admins & recruiters, pick the new owner, and confirm.
• From that moment, every assessment-completed, hired, and rejected email for the job lands in their inbox instead — no more forwarding by hand.

Candidates often open an invite wondering how to even start. Now you can answer that right inside the email — without building it by hand.

• Type / in the email body, or use the insert menu, and you'll find a new Suggested group at the bottom.
• Three ready-made, on-brand blocks: a Support video thumbnail, a Getting-started banner, and a short Help / support link.
• Each drops in complete — thumbnail, link, and copy already filled — and you can edit it afterwards like any other block.

Candidates used to skim the start screen, hit Begin, and get tripped up mid-test — by the timer, the camera, or thinking that finishing one question ended the whole thing. Now they walk through it first.

• A short, stepped intro replaces the old wall of text: what you'll do → timing → camera & your space → ready, one idea per screen.
• The timing step shows the real numbers and makes it unmistakable whether each question is timed on its own or one clock runs the whole way.
• It states plainly that the camera and mic stay on and record the whole time, and asks the candidate to find a quiet, well-lit spot.
• The camera turns on quietly in the background while they read, so there's no waiting around when they reach Begin.

Some screens are dense enough that a first-timer just stares. Now the rest of the screen dims, one thing lights up, and a plain-language card explains it — one step at a time.

• In the coding assessment, a walkthrough covers the numbered question boxes, the green ✓ that means solved, the difference between Run and Submit, and that you submit each question on its own.
• In the email template editor, a walkthrough covers building with blocks, personalizing with variables, sending yourself a test, and the toggle between Goodfit's polished default and your custom version.
• Each shows up on its own the first time, and a ? button replays it whenever someone wants a refresher — it never replaces the option to reach a human.

The goodfit API could already drive your hiring end-to-end — but you had to keep asking it what changed. Now it can tell you the moment something happens.

• Subscribe a URL to the events you care about — 16 of them, spanning jobs, applications, assessments, candidates, and prescreening.
• When a matching event fires — a candidate finishes an assessment, gets hired, passes prescreening — goodfit POSTs a signed JSON payload to your endpoint within seconds.
• Delivery is durable: it's queued, retried with backoff, and every attempt is logged, so you can replay anything your system missed while it was down.
• Manage it all over the same x-api-key API you already use — create subscriptions, page the delivery log, or fire a test payload at any URL before going live.

// POST → your endpoint, on every matching event
{
  "id": "9f2c…", // unique per event — dedupe on this
  "event": "assessment.evaluated",
  "createdAt": "2026-06-15T12:00:00.000Z",
  "organizationId": "Qu1Dl…",
  "data": {
    "applicationId": "…",
    "assessmentType": "voice",
    "score": 8,
    "maxScore": 10,
  },
}
// signed with your secret in the X-Webhook-Signature header — verify before trusting

The strongest resume matches in your workspace now appear right where you work — ready to invite in one move.

• Flip on ✨ Show AI candidate suggestions and your closest resume matches appear under Prospects, each with a fit score.
• Drag one onto Invitations or hit Invite Candidate — send the interview invite.

Mentions, completions, and invites used to live only in your inbox. Now they also live as in-app notifications.

• Grouped Today / This Week / Earlier.
• It pings you when a teammate @mentions you, a candidate finishes their pipeline, or you're invited to an organization.
• A gentle pop-up for the moments worth interrupting you; a silent badge for the rest.

Two upgrades on a candidate's feedback panel: pull a colleague in, and see where your interviewers landed without doing the math.

• Type @ in a comment to tag a teammate.
• A new Team verdict header turns every scorecard into one read: a lean like Leaning Yes, a bar with a segment per interviewer, and a Split badge when they disagree.

Last release the talent pool learned to read plain English. Now a single job's applicants board does too.

• Hit ✨ AI and type something like proficient in Malayalam, applied within the last week.
• It builds the matching filters; you tick the right ones and apply.
• Can't express something? It tells you in plain words — never a silent wrong guess.

Add subjective questions for skill based assessments. Now each one carries its own grading method, so you decide what "correct" means.

• Pick how a Subjective question scores: keyword, exact, or pattern match, AI grading against an ideal answer, manual review, or collect-only.
• The match methods score instantly, no AI; manual review lets you read and score it yourself on the report.

A candidate runs, fails, edits, submits again. Now you can see what they changed to finally pass.

• A Code / Diff toggle diffs any submission against the Previous, First, or Accepted one — added and removed lines called out.
• Unchanged code collapses away, so the change is front and centre.

New to Goodfit? A small checklist in the sidebar shows the first four steps — then bows out.

• Create a job → invite a candidate → invite a teammate → hire — each step links to where it happens.
• It ticks itself off as you work; no wizard, no tour.
• Dismiss it whenever, and it vanishes on its own once all four are done.

We’ve launched Goodfit MCP — now you can connect Goodfit directly using the Goodfit MCP itself, or set it up within Claude.

• Generate a key and copy a ready-to-run command for Claude Code or a config block for Claude Desktop.
• Paste it, and Claude can read and act on your jobs, applications, and candidates.

Last month the talent pool let you narrow on three things. Now it answers the questions you actually have — "who scored at least B2 on our English test, passed the JavaScript assessment, has React on their resume, and applied in the last month" — as one search.

• A new Filters dialog with sections down the side — Jobs, Application, Identity, Skills, Experience, Education, Location, Assessments, and Quality — each showing how many filters you've set so you never lose your place.
• Filter on every assessment you run, by name: set a score range (70–100%), or for language tests a minimum level (English at least B2). The list grows with your assessment library — language tests included, which the old filter couldn't even show.
• Slice by what's genuinely on a resume — a specific skill, the candidate's current employer, a past role, school, degree, city, country, years of experience — matched against the real details, not a stray word buried in a job description.
• Add quality and integrity signals: resume score, AI-writing percentage, and proctoring flags. Or paste a list of emails to pull up exactly those people.
• Stack as many as you like and we get the logic right — "B2 English and 70% coding" means one candidate who did both on the same application, never two unrelated results stitched together.

Not sure which filter says what you mean? Type it as a sentence and let Goodfit set the filters for you.

• Click the ✨ AI button next to Filters and type something like designers from Mumbai fluent in Marathi.
• It proposes the matching filters — Role contains "Designer", City contains "Mumbai", Skill contains "Marathi" — and you tick the ones that are right before applying.
• It never guesses silently: anything it can't translate lands in a Couldn't handle list with a plain reason, so your results always mean exactly what you asked.
• Confident matches come pre-ticked; the iffy ones — a state typed where we only have a city field, say — come unticked, so you stay in control.

Your public careers page used to come in exactly one layout. Now pick from seven branded templates and see your page redraw the moment you choose.

• A new Design panel in Settings → Careers with seven templates — Classic, Minimal, Bold, Editorial, Warm, Dark Matters, and Radiant Play — each one a live little preview of your own page.
• A full preview sits right beside your edits and updates on every keystroke — no more saving, publishing, and opening a new tab just to see a change.
• Flip between Desktop and Mobile to see exactly what candidates get on a phone.
• Light or dark, sharper or rounder corners, tighter or roomier spacing — every option is pre-designed, so the page always looks intentional. Your brand color flows in as the accent automatically — no color-picking, no fonts to upload, nothing to get wrong.
• Already published a careers page? It keeps its current look until you pick a template.

Built a job's hiring stages and need a manager's sign-off? Send them a link that walks every screen a candidate will see — no screenshare, no taking the assessments themselves.

• A new Preview as candidate button in the pipeline editor — and Preview before publishing on the publish step — opens a read-only walkthrough.
• Step through the whole flow in order: the application form, each assessment, the voice interview, the coding problems — exactly as a candidate would, with nothing submittable.
• Every stage carries a Recruiter settings strip — pass score, what happens automatically, whether the stage is mandatory — so the manager can vet the questions and the thresholds in one place.
• Hit Share to copy a link; teammates land on the saved pipeline, making for a clean Slack handoff.
• Reorder or tweak stages without saving and the preview reflects your unsaved draft.

Every job now has its own intake email address — candidates apply just by emailing their resume.

• Turn the address on for a job and share it on LinkedIn, X, WhatsApp, or anywhere else.
• A candidate emails their CV from anywhere; we parse the resume, score it against the JD, and create a prospect card with a 📧 Email Inbound badge.
• Click View email on the application to read the original message and download every attachment.

Until now you had four built-in emails (invitation, assessment, rejection, hired) and nothing else. You can now write your own — interview confirmations, take-home briefs, offer letters, polite nudges, reference-check requests — and send them straight from a job's pipeline.

• Settings → Email Templates → + New template. Author a subject and body with click-to-insert variables for candidate name, job title, company, and the sender.
• Every template renders inside your branded layout — your logo on top, your brand color on every button — so what you preview in-product is exactly what the candidate sees in their inbox.
• Send from the kanban: select candidates on a job's board, open the Send email dialog, pick a template, hit send. One email, branded, audit-logged, done — no copy-paste from Gmail.
• Attach real files to any template. Flip Requires attachment on (e.g. "Offer Letter" or "Take-home brief"), and the kanban will only let you send it to one candidate at a time up to 3 files, 3 MB each: PDF, DOCX, PNG, JPG.
• The candidate gets a branded email with a real paperclip. The application's activity timeline records the template name and the files that went out — so "what did we send Bob three months ago?" stops being a mystery.

Don't want to keep refreshing the pipeline? Let the job email you the moment a candidate completes.

• On the job edit page, flip Settings → Email on assessment completion (off by default).
• When on, the person who created the job gets an email per completed attempts. View full report button straight to the application.

Turn on Resume-aware AI Interviewer on a job and the voice bot reads each candidate's resume before the call — then uses it during the interview.

• It anchors on what's on the CV ("Your resume shows ~6y with React — how do you approach state management?")

Reviewing a proctored attempt used to mean flipping between two separate recordings — one of the candidate, one of what they were doing — and trying to line up the timing by eye. Now they play together.

• When a candidate pauses for forty seconds, you can see at a glance whether they were thinking, looking away, or no longer at their desk — without ever leaving the page.
• The candidate's video sits as a small circle in the corner of their work. Hover it to enlarge; click it to open in a new tab.

Evaluations on the application's Evaluation tab used to rate against the job-level skill list — which often drifted from what the voice interview actually assessed. Now the scorecard mirrors the interview.

• The skills you see on the scorecard are pulled straight from your voice interview skills for that stage. If the job has more than one voice round, we combine them and drop duplicates.
• When there's no voice round, the scorecard falls back to the job's skill list and lets you add a custom skill row.
• Your teammates' scores are hidden by default — no more anchoring on "2 Yes, 1 No" before you've written your own. Click View team evaluations when you're ready, Hide team evaluations to collapse them again.
• Forget to rate a skill or pick a recommendation? You get a clear message saying what's missing, instead of a silently disabled button.

The Talent Pool gets a new Filters button next to the search / job / status / sort row, with three deeper ways to narrow down candidates.

• When they applied — pick a date range with the two-month calendar.
• Resume mentions — type a keyword (e.g. react, adobe, gst). We search the parsed contents of every resume. Each matching candidate shows the surrounding snippet right under their email with the keyword bolded — so you can tell real Java from JavaScript-only at a glance.
• Assessments completed — pick from a list of your organisation's actual assessments by name (not generic types). Selecting one means "this candidate has finished and been evaluated on this specific assessment."
• Filters live in the page address — bookmark a useful slice, share it with a teammate, reload the page and it's still there.

The Download Applications button on a job's kanban now respects your active filters. The CSV contains exactly what's on screen — not the whole job.

• No filters active? You'll get the full applications list, as before.
• Filters active? Hover the download button and it tells you exactly how many applications are about to leave with you — Download 12 filtered applications — and the CSV mirrors that.
• If your filters narrow the board to zero candidates, the button stays disabled — never a silent full-job export when you thought you were narrowing.

• Hover the Download CSV button on Overview, Pipeline Health, Sourcing Analytics, or Team Activity and you'll now see Download filtered data — a small, clear confirmation before you click.
• Narrow to "Q1 2026, recruiter X" and the CSV is that slice, every time.

Creating or editing a job used to throw fourteen open cards at you in one wall — overwhelming, hard to scan, and easy to lose your place in. We've cleaned that up.

• The most important sections are open when you arrive — job status, visibility, basics, job description — and everything else sits one click away, in order, ready when you need it.
• No more scrolling through settings you don't care about to reach the one you do.

The voice interview's opening and closing lines can now address each candidate by name.

• In the Ice breaker and Closing remarks fields, type @ to insert a chip: Candidate's name, Company name, or Job title.

• At call time the bot speaks the finished line — "Hello Alice, thanks for your time interviewing for the Senior Engineer role at Acme."

  

Choosing coding questions is now a full-screen, browse-and-filter experience. The prebuilt coding bank is now 430 coding problems across 10 languages and 95 SQL problems.

• Topics span data structures, algorithms, backend logic, and real SQL analytics (revenue deltas, org hierarchy, quartile buckets).
• Filter rail down the left — Type, Difficulty, and Topics, each with a live count.
• Search, sort (default / title / easy→hard / shortest).
• Click a question to preview it in a side pane before adding.
• Multi-select with a review strip at the bottom — see everything queued, remove any, then "Add N questions".
• Curated bundles — five ready-made starter packs (Coding Fundamentals, Quick 30-min Screen, Backend Engineer Screen, Data / SQL Analyst, Algorithms Deep-Dive) as one-click cards.

You can now pull single questions from a library of ~10k prebuilt skill questions — not just adopt a whole assessment.

• Browse and filter by family, skill, difficulty and type, with full-text search.
• Preview each question — including its image or audio media — with the options and the rationale.
• Pick questions one at a time and mix them with your own, or add a curated pack (organised into 16 family categories → packs → questions).

A skill assessment can now serve a random handful of its questions to each candidate — so the same questions don't leak between candidates.

• Turn on Random subset in the assessment settings and set how many questions to show.

Language-proficiency tests are no longer English-only, and you can read every question before you assign one.

• Available languages now include:
  • English
  • Hindi (हिंदी)
  • Telugu (తెలుగు)
  • Tamil (தமிழ்)
  • Gujarati (ગુજરાતી)
  • Malayalam (മലയാളം)
  • Marathi (मराठी)
  • Kannada (ಕನ್ನಡ)
  • Odia (ଓଡ଼ିଆ)
• Hit View questions on any language to open a full read-only preview — reading passages, audio clips, every option, and the answer key — exactly what the candidate will see.

When you write your own voice question, you no longer have to hand-write the scoring guide.

• A Generate with AI button now sits next to each ideal-response slot — both the primary question and the follow-up.

Found good people in your talent pool? Drop a batch straight onto a job — already scored.

• Select candidates, click Add as Prospect, pick one target job.
• Each candidate is scored against that job's description and added as a prospect with the score visible.
• A summary toast tells you what happened — "3 added, 1 already on this job, 1 skipped (no resume)".

Candidate email is now durable and observable — and you find out fast when your mail is misconfigured.

• If your custom SMTP is broken, Email Settings shows a red "Delivery failing" banner with the exact provider error and provider-specific fixes (including Microsoft 365).
• Application cards and the activity timeline now show per-message delivery status alongside stage moves.

Application forms can now grade candidates with points — not just pass / reject.

• New question type: Scored multiple choice. Each answer carries 0–100 points (better answers = more points).
• Set a Minimum total points under the form's settings — anyone below it is auto-rejected, the same way knockout questions already work.
• Live helper shows the best possible total as you build, so the minimum is never set higher than what's achievable.
• Candidates only see the answers — never the points.
• Recruiters see the score on the report — total, max, and a green +points chip next to the answer the candidate picked.

Application forms can now branch — a question can be set to appear only when an earlier answer matches.

• On any question, open Advanced → Conditional logic and pick an earlier question + the answer that should trigger it.
• Works with: yes/no, dropdown, scored multiple choice, multi-select, number, and date questions.
• Hidden questions are completely ignored — not asked, not required, and they can't reject anyone.
• Candidates see a clean, shorter form tailored to their earlier answers.

Picking a psychometric stage is now a visual choice — and you can read every question before committing.

• All 5 frameworks now appear as cards with number of traits measured and a description. With an ability to view every question, the trait it measures, and the response scale used

Voice interview reports now tell you, whether the candidate's connection held up during the call.

• A small Connection tag appears next to "Time Taken" on every voice attempt — Spotty Connection (amber) or Bad Connection (red).
• Hover the tag for a narrative explanation: what the candidate experienced ("about half the call's voice data was lost in transit"), what network they were on, and — for bad connections — a clear recommendation.
• Clean calls show nothing — the tag only appears when it's worth your attention.

Spotted a wrong score, a broken transcript, a missing recording, or anything that doesn't add up? Flag it in one click.

• New 🚩 Flag icon at the right of the report's results strip — on both assessment and prescreening reports.
• Click it, type a short note, hit Flag.
• The flag goes straight to the Goodfit team's internal channel — so we can investigate without going back and forth.

Every important page now has a question-mark icon next to its title — click it to open the matching help article.

• Icon appears on Create Job, Email Settings, Careers Page, Members & Permissions, Employee Referrals, and more (we'll keep adding).

  

Pay-as-you-go billing now powers every new company on Goodfit.

• Every organization created on or after Saturday, 25 April 2026 is on metered billing
• 20 free credits to start with — no card required
• One credit per assessment (coding, voice, phone, psychometric, skills)
• Top-ups available at ₹100 + GST per credit, minimum 100 credits per purchase

See exactly where candidates drop off in your hiring funnel.

• Pipeline Health view under Insights

• Waterfall chart showing how many candidates moved from one stage to the next

• Final conversion to Hired or Rejected at a glance

• Spot the stage that's leaking the most candidates

  

Invitations now have a shelf life.

• Default expiry: 3 months from the day the invitation is sent
• Configurable per job — minimum 1 hour, maximum 1 year
• Once a candidate starts the interview, the expiry no longer applies — they can finish at their own pace
• Invitation emails now show the deadline ("Please respond by DD MMM YYYY")
• Expired invitations show a red Expired pill on the Kanban tile
• Re-inviting a candidate resets the clock.

AI Interview is now split into two clear products at every step.

• AI Web Interview — candidate joins from a browser link
• AI Phone Interview — our bot calls the candidate directly on their phone

Phone interview rules

• A phone-interview job can only have phone-interview stages — no other stages can be added before or after
• No application form can be attached to a phone-interview job
• For public jobs, candidates pick the date and time they want to be called — right on the job page

Voice interviews can now go deeper on each competency.

• One voice interview can have multiple competencies
• Each competency can have multiple primary questions plus their follow-up questions
• Each primary and follow-up question carries its own ideal response for the AI to score against

Reports are now private by default. Share deliberately, share safely.

• Reports can no longer be viewed publicly unless you generate a Share URL
• Choose who can see a shared report:
  • Specific people by email
  • Specific company domains (e.g. anyone at @yourcompany.com)
  • Fully public link
• Toggle visibility on or off any time

Per Share URL, you can configure:

• General access — who's allowed to open the link
• Link expiry — set a date when the link stops working (or "No expiry")
• Hide from viewers — pick what to redact from the report:
  • Hide candidate PII — name, email, phone, location, photo
  • Hide screening form responses
  • Hide internal evaluations (scorecards, hiring-manager notes)

Useful for blind reviews, external interview panels, or sharing with clients without leaking your private notes.

Candidates no longer see a harsh rejection screen when there's nothing left to evaluate.

• If a candidate falls below the cutoff at the last stage — or if everything left is a psychometric, they finish the pipeline normally and see the same completion screen as anyone else

• Recruiters still see the truth on the Kanban board — the candidate still lands in Poor Fits with the real score

• No change for failures earlier in the job stages pipeline — those still show the rejection screen as before

A new per-assessment toggle: Disable Editor Locking.

• Off (default) — strict editor: no copy/paste, no right-click, no autocomplete. Same as before.
• On — relaxed editor: copy/paste, IntelliSense, right-click, command palette all work normally

Useful for when the standard lockdown feels too aggressive.

Recruiters can now pick the voice (and accent) the AI voice interviewer uses, per interview language — giving teams finer control over how interviews sound for different candidate cohorts.

• English — 12 voice options across US / UK / India / AU
• Indic languages — 6 voice options per language

Jobs can now carry structured compensation ranges at create or edit time.

• Salary range with minimum and maximum values
• Equity range with minimum and maximum values
• Improves listing quality when Google Jobs distribution is enabled

Application and pre-screening are now a single form flow. Candidates can submit the form even on a private job — including when they haven't been invited — as long as a form is attached.

Form outcome behavior hierarchy

Auto-Advance from Prospects

Prospects who completed the application form (passed or held for review, but not failed) now show an Auto Advance button on their Kanban card. One click skips the invitation stage and moves the candidate straight into the first assessment — assessment invites (email + WhatsApp if enabled) are sent automatically.

Recruiters who haven't completed onboarding now see a lightweight onboarding screen. The step is optional and can be skipped at any time.

Collects:

• Name
• Phone number
• Designation
• Team size
• Goals

Recruiters can now revoke a candidate from the Prospects or Invitation stage directly on the Kanban board.

• For private jobs, a revoked candidate can no longer start the interview.
• Revoking is a board-level action — it does not recall or delete any email or WhatsApp invitation already sent to the candidate.

Recruiters can filter applications on the Kanban board using structured resume-evaluation fields:

• Worked At
• Current Company
• Past Role
• Studied At
• Degree
• Field of Study
• Skill
• Certification
• Language
• Country
• City
• Years of Experience
• Avg Years / Company
• Frequent Job Hopper

End-to-end referral pipeline to streamline candidate sourcing through your team.

• Whitelist specific company domains to allow internal referrals
• Enable referrals at the job level
• Employees can share and refer candidates through the Goodfit job board
• Referred candidates automatically land in the Invitation stage on the Kanban board
• Referral invitation emails are sent with an application link
• Filter and track referred candidates directly on the Kanban board

Helps spot if someone might not be honestly answering voice interview questions by checking if their spoken words match their lip movements in video recordings.

After you finish a voice interview, the tool reviews the recording, checks when the candidate is speaking, and makes sure their lips and voice match up.

• Flags moments where something seems off between lips and voice
• Shows if the candidate's video and audio are in sync for different parts of the interview
• Gives a simple result: Genuine, Borderline, or Suspicious, so you know if you need to look closer

• Bulk download interview reports directly from the Kanban board
• Reports are delivered via email as a ZIP of PDFs — no waiting on the UI

Structured evaluation system for hiring teams.

• Rate candidates on a scale from Strong No → Strong Yes
• Evaluate against job-specific skills
• Add notes per skill
• Add a final overall note for the candidate

• Embed jobs directly on external websites
• Job distribution via Google Jobs (listings may take 2–3 days to appear after publishing)
• Choose distribution targets per job: company careers page, Goodfit careers page, Google Jobs

AI now generates ideal responses (scored 1–10) during question creation for both primary and follow-up questions, giving evaluators a reference to compare against.

Microsoft Outlook authentication is now supported alongside existing providers.

Inviting candidates one at a time is gone.

• Paste or upload a list of emails and invite the lot in one go
• Upload a folder of resumes; each becomes a candidate
• Or add candidates by hand, one row at a time
• A confirmation step before anything is sent, so a bad paste doesn't reach 200 inboxes
• Invitations go out by email and WhatsApp

• Resumes uploaded to a job land as prospects, parsed into name, contact and experience
• Each is scored against the job, with the reasoning attached
• Preview the resume itself without downloading it
• Duplicates are flagged rather than quietly added twice

• Every assessment a candidate took, on one page, with scores broken down by section
• A timeline of what happened and when: invited, started, submitted, evaluated
• Export the board, or any filtered slice of it, as CSV

• Choose whether the clock runs per question or across the whole assessment
• Weight questions so the ones that matter count for more
• Candidates get a start screen explaining what's coming before the timer starts
• Candidates pick the language they'll be assessed in

• A liveness check before an assessment begins, so the person who starts it is the person who's there
• Written answers are checked for signs they weren't written by the candidate

• Every change to a job is recorded: who made it, what changed, when
• Seniority, tools and skills added to the job details
• Filter and sort the jobs list

• A personal settings page, with an avatar you can upload
• Team management in organisation settings, with roles per member

• A full editor for the candidate: problem statement, starter code, and a console
• Write test cases once; every submission is run against them and scored automatically
• A per-attempt report showing what was submitted, which tests passed, and the score
• A countdown the candidate can see, and a submission that fires on its own when time runs out

• A live spoken interview the candidate joins from the browser
• Microphone and camera check before the call, with device pickers and a sample recording to test against
• The call is recorded and transcribed as it happens
• Answers are scored afterwards against the question, with the transcript kept alongside

• Pick a framework, set the response scale, and assign it to a pipeline stage
• Responses are evaluated and returned as a profile, not just a number

• Questions that ask for a recorded video answer
• Multiple-choice questions with options, scored on submission

• Resumes are read and scored against the job before anyone opens them
• The score and the reasoning sit on the candidate's card

• Suspicious behaviour during an assessment is recorded against the attempt
• The browser is hardened for the duration of an assessment

• Branded templates for invitations, magic links and organisation invites
• Sent from Goodfit, tracked, and rendered the same in every inbox

Every applicant to a job, laid out by the stage they're in.

• Drag a candidate from one stage to the next
• Search by name, filter by stage, score or date, and combine filters with AND or OR
• Move a candidate's status without leaving their card
• Bulk actions across a selection, and a count on every column

• Add, reorder and rename stages, including custom stages with a colour of their own
• Attach an assessment to any stage
• Changing a live pipeline no longer strands the candidates already in it: choose whether they stay on the old version, are rejected, or are moved forward

• Build an assessment once and reuse it across jobs
• Questions organised by skill and microskill, drawn from a skills taxonomy that spans engineering, sales, operations and more
• Set the duration, and preview the assessment as the candidate will see it

• A rich text editor for the description, with formatting, lists and links
• Drop images straight into the description

• Invite candidates by email from the job itself
• Invited candidates land on the board, ready to move

Goodfit is now something a team uses, not one person.

• Create an organisation and switch between the ones you belong to
• Invite teammates by email; they join with a role that decides what they can reach
• An organisation settings page for the details and the member list

• Sign in with a link sent to your inbox, no password to remember
• A waiting screen that tells you the mail is on its way, with a jump straight to Gmail or Outlook

• A guided create flow that prefills what it can
• Locations, visibility and the rest of the details, editable after publishing
• Each job carries a pipeline, and each pipeline stage can hold an assessment

• Every job gets a public page candidates can read and apply from
• Mark a job public or keep it private to people you invite

• Email and password sign-in, with social providers alongside
• Sessions that persist, sign-out, and dashboard pages that stay private

• Sidebar navigation, breadcrumbs that track the page you're on, and a light/dark toggle
• One interface library behind every screen, so the product looks the same wherever you land

• Create a job and edit its details