MyRecruiter AI — The Introduction Engine
A lean Express /ask endpoint proxying to OpenAI — the seed of a production introduction engine that writes bespoke recruiter introductions, individually metered and billed per candidate.
Back Story
Recruiters spend hours writing candidate introductions — those carefully worded summaries that pitch a candidate's background, skills, and fit to a hiring manager. It's high-stakes, repetitive, and surprisingly hard to do well. MyRecruiter AI was built to automate that work: send a candidate profile to a single /ask endpoint and get back a polished, role-aware introduction ready to paste into an email or CRM.
The long-term vision is a separately-metered runtime product — each introduction costs a small, tracked token spend, billed per candidate — so the economics are transparent and the service pays for itself.
The Build
- Built a minimal Express.js application with a single POST /ask endpoint that forwards a structured prompt to OpenAI's chat completion API.
- The endpoint accepts a candidate JSON payload (name, title, experience summary, key achievements, target role) and returns a formatted introduction paragraph.
- Designed for deployment as a standalone service — its own repo, its own runtime, its own API key — not coupled to any WordPress site.
- Currently lives on HostGator shared hosting (in the myrecruiter monorepo's ai-service/ directory) pending split into its own dedicated repo.
Headline Tasks
- Wired Express routing, request validation, and OpenAI chat completion call.
- Structured the prompt template — system message defines the recruiter's voice and introduction format; user message carries candidate-specific detail.
- Defined the build order: C (basic intro engine) → B (candidate data enrichment) → A (full product with billing and CRM integration).
- Currently blocked on a valid OpenAI API key — key in .env is a placeholder, awaiting production credentials.