Jobraker Recruiters
Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

Jobraker Recruiters is a local-first recruiting workspace for founders, solo recruiters and lean hiring teams. Hiring work is scattered across job briefs, candidate profiles, email, calendars and notes. The project turns that fragmented work into a persistent, inspectable workflow that can continue between conversations. The implementation extends an existing Electron desktop application. A recruiter selects a role, sets its requirements and delegates a bounded mission. Specialist tasks for sourcing, evidence review, drafting and scheduling are linked to the same role and candidate records. Shared memory, durable task state, budgets, permissions and provider receipts make the work reviewable. Scheduled reply checks and bounded Gmail follow-ups are implemented for approved campaigns. The Meta integration targets Muse Spark 1.3 for planning and tool use. Candidate assessments are tied to explicit requirements and source evidence. External messages and invitations require a connected provider and valid permission or exact approval. Final hiring decisions stay with the recruiter. The application preserves Home, Roles, Candidates, Pipeline, Sourcing and Analytics views. Hiring context stays in inspectable local records and Markdown notes. Windows and Linux X11 computer-use adapters are implemented with selected-application scope and reviewed inputs. The Meta recruiting source is on branch feat/meta-autonomous-recruiting-os in the linked repository. Its verification record reports passing core, renderer and existing-action tests, builds and lint checks. A real read-only GitHub lookup and isolated Linux X11 interaction were verified. Meta inference and write-oriented recruiting providers were tested with controlled fixtures. Live Meta execution and real Gmail/Calendar actions remain unverified due to account and runtime setup issues. These implementation and test results are not a claim of a completed live AI recruiting campaign.