Hyrvanta is a privacy-conscious career workspace that helps job seekers discover verifiable opportunities, assess job fit, manage applications, and tailor resumes without inventing experience.
Release status: Portfolio-ready working release. The interface uses clearly labeled preview records until production job sources, identity, and supporting services are configured.
Most job-search tools optimize for volume. Hyrvanta is designed around clarity, evidence, and user control:
- Evaluate opportunities clearly with source, freshness, availability, distance, and fit shown as separate signals.
- Build resumes from evidence by linking material claims to records in the Career Vault.
- Keep the user in control—Hyrvanta never submits an application or changes its stage automatically.
- Protect career data with owner-scoped records, authenticated APIs, private object storage, and explicit privacy controls.
- Stay honest about uncertainty by labeling demonstrations, missing data, and unavailable integrations instead of simulating results.
| Workspace | What it provides |
|---|---|
| Dashboard | Guided setup, recommended next actions, and an at-a-glance application pipeline. |
| Discover | Role, city, radius, work-setup, employment-type, and experience-level search with separate Discover, Analyzed, and Saved Searches views. |
| Applications | User-controlled tracking for saved, applied, interview, offer, and rejected stages. |
| Resumes | Private PDF, DOC, and DOCX upload, stored-file management, version history, and evidence-aware tailoring review. |
| Career Vault | Structured profile, experience, education, project, certification, achievement, and skill evidence. |
| Insights | Evidence-based job-requirement and pipeline summaries that disclose sample size and date range. |
| FAQ & Assistant | Accessible help content and deterministic, grounded guidance without invented listings or career history. |
| Profile & Data | Search preferences, permissions, extension status, notification controls, and data-management actions. |
- Greenhouse and Lever employer-board adapters
- Optional Jooble adapter
- Normalization and deduplication across sources
- Verification, freshness, URL-safety, retry, timeout, and Haversine-distance rules
- Protected synchronization endpoint with source-isolated failure handling
- Manual job capture and deterministic analysis
- Authenticated resume upload to private R2 object storage
- Owner-checked preview, download, and deletion
- D1 metadata and version tracking
- Career Vault evidence references for material resume changes
- Explicit separation between professional experience, projects, labs, education, and certifications
- Responsive React interface with direct, server-renderable application routes
- Normalized D1 schema and versioned Drizzle migrations
- Authenticated, owner-scoped APIs with prepared queries
- Manifest V3 browser capture extension
- Transport-independent MCP tool registry
- Accessible navigation, forms, FAQ interactions, status messaging, and reduced-motion support
Generated resume text is considered valid only when each material claim has one or more traceable Career Vault evidence references. Hyrvanta does not convert coursework or lab practice into professional experience, infer sensitive application answers, present preview listings as current vacancies, or submit applications automatically.
- Node.js 22.13 or newer (below Node 26)
- npm
git clone https://github.com/ailevated-randy-alonzo/hyrvanta.git
cd hyrvanta
npm install
npm run devOpen http://localhost:3000/dashboard.
The local interface remains usable with preview data when hosted identity, employer feeds, or other production integrations are not configured. Features that require those services identify the missing dependency in the UI.
npm run typecheck
npm run lint
npm testnpm test runs the domain and route tests, browser-like React interaction tests, a production build, and rendered-HTML checks. To create a migration after changing the D1 schema, run:
npm run db:generateReact workspaces and public pages
|
authenticated routes
|
Cloudflare Worker runtime
| |
D1 records R2 resume files
|
job-source adapters + MCP tool registry + browser extension
Important directories:
app/ Application routes, public pages, and APIs
browser-extension/ Manifest V3 job-capture extension
db/ Drizzle schema and D1 access
docs/ Architecture and engineering decisions
drizzle/ Versioned database migrations
lib/ Domain rules, matching, sources, and shared services
mcp/ Transport-independent MCP tool definitions
tests/ Domain, route, UI, extension, and render checks
worker/ Cloudflare Worker entry point
A production deployment requires configured identity, D1 and R2 bindings, employer-board identifiers, monitoring, and secrets. Optional services such as email delivery, document-text extraction, production AI generation, and browser-extension pairing are intentionally shown as unavailable until they are configured.
Never commit .env files, credentials, production binding identifiers, resume contents, or raw captured job-page HTML. See the deployment checklist and security model before publishing a production environment.
- Architecture
- Current implementation
- Accessibility
- Browser extension
- Data model
- Deployment
- Design system
- Development
- Matching engine
- MCP registry
- Resume engine
- Security and privacy
- Evidence before assertion — show where a claim or recommendation came from.
- Decision support before automation — help users decide; do not act in their name.
- Privacy before convenience — minimize access and keep sensitive data owner-scoped.
- Honest states before polished fiction — identify preview, unavailable, incomplete, and unverified states plainly.
- Accessibility as a requirement — support keyboard use, visible focus, readable contrast, semantic structure, and reduced motion.
Built as a portfolio-scale demonstration of trustworthy career technology, full-stack product engineering, and user-centered interface design.