goodfitdocs
HelpChangelogPlatformAPI docsDashboard
Release notes
v0.0.13

June Week 2

Find your best-matching candidates as ready-to-invite cards on the board, catch every mention, completion, and invite in a new notification bell, @mention teammates and read your panel's verdict at a glance, filter applicants in plain English, choose how each written answer is graded, diff a candidate's coding attempts, and connect your jobs and candidates to Claude in one paste.

All releases →

Hiring

Meet your best-matching candidates right on the board

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.
A job's Prospects column with a dashed AI Suggestions divider and faded candidate cards below — each with an avatar, name and headline, an open-resume arrow, a circular match-percent ring, and a full-width invite button
Your closest matches, surfaced under Prospects — invite any of them in one move.
The job board toolbar with edit, download, PDF, a highlighted sparkle button, and plus; a tooltip over the sparkle reads Show AI candidate suggestions
One sparkle button turns it on — and it's off until you ask for it.

Notifications

Notification bell, so nothing slips past you

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.
A bell with a red unread badge in the navbar, and an open panel headed Notifications with a Mark all read action, a Today bucket, an org-invite row, a mention row with an unread dot, and a See all footer
One glance tells you what happened while you were away.
A notification pop-up reading You're invited to Homebrew Coffee with the org name as a subline and a blue Join organization button
The moments worth a pop-up come with a button straight to the action.

Hiring

Loop in a teammate, and read your panel's verdict at a glance

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.
A comment composer with the @-mention autocomplete open — typing @ja surfaces a teammate with avatar, name, and email in a dropdown above the box
Type @ and pick a teammate — they get an email that lands them on the candidate.
A Team verdict header — a Leaning Yes label with a Split badge, a per-interviewer pastel segment bar, a 4 Yes / 1 Neutral / 1 No summary, per-skill averages, and a breakdown with the viewer's own row marked You
Where your panel landed, in one line — lean, split, and a segment per interviewer.

Hiring

Filter a job's applicants by typing what you mean

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.
The AI search popover over a job's applicants board — the query proficient in malayalam, applied within the last week resolved to two ticked suggestions with a Dismiss and Apply footer
Type a sentence, review the filters it suggests, apply the ones you want.

Assessments

Choose exactly how each written answer is graded

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.
The Evaluation dropdown on a Subjective question, open — six options: Keyword match, Exact match, Pattern match, AI grading, Manual review, Collect only
Six ways to grade a written answer, chosen per question.
A Subjective question in the builder with Evaluation set to Keyword match, a keyword tag input, and Require all keywords and Case sensitive checkboxes
Each method reveals just its own settings.

Assessments

See what changed between a candidate's coding attempts

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.
The coding submission viewer in Diff mode — a vs Run header with green added and red removed line counts, a Code / Diff toggle, an inline diff of a changed return line, a connected pass/runtime/memory strip, and two failed test cases with Expected and Got
The diff puts the change front and centre.

Getting started

A quiet 'Get started' checklist for new workspaces

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.
A minimal Get started card in the sidebar footer with a 3 of 4 counter and a dismiss X, above four rows — three completed and struck through, one open for Hire your first candidate
A periphery nudge through the first four steps — it fills in as you go, then retires.

Developers

Goodfit MCP is here! Connect Goodfit to Claude

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.
A Connect to Claude reference card with the Claude mark, a segmented Claude Code / Claude Desktop tab strip, and the generated command in a monospace block with a corner copy button
The setup recipe lives right on the API Keys page — copy, paste, done.
Claude Code after running the generated command — the goodfit server connected and listing its tools for jobs, applications, and candidates
After one paste, Claude sees your whole Goodfit toolset.

June Week 3

Score how well a candidate actually speaks the interview's language, hand a job and its notifications to the right teammate, drop a ready-made support video or getting-started banner into your invitation emails, walk candidates through what to expect before every assessment, light up the screens people get stuck on with guided walkthroughs, and stream your hiring events to your own systems in real time.

June Week 1

Slice the talent pool by every signal you collect — assessment scores, language levels, resume details, and quality flags — or just describe who you want in plain English and let AI build the filters, style your public careers page from a gallery of templates with a live desktop-and-mobile preview, and hand a hiring manager a shareable walkthrough of every screen a candidate will see.