2) Edit precision + job-fit diagnostics

Combine Form + YAML editing with AI scoring, keyword-gap analysis, and selected-company targeting from external companies metadata so every change can be checked for recruiter relevance and missing signal.

3) Presentation polish and profile quality

Review available CV templates side-by-side, then finalize profile-photo quality using Photo Booth analysis and recommendations as part of the full self-hosted CV authoring workflow.

1.0.2 Release Scope

Version 1.0.2 is the current public release with a stable user-facing workflow:

• Print Room (preview + export)
• Editor (Form/YAML)
• Editor Form View includes collapsible nested containers (collapsed by default for deep structures) with compact summary metadata for faster navigation.
• Experience entries in Form View are labeled with role-first titles and period/company subtitles instead of generic numbered labels.
• Settings (OpenRouter login, Analysis Model, Image Generation Model for future use, base URL, credit status, and per-check cost estimates)
• AI analysis targeting with metadata-source selection, multi-company checkbox selection, and inline Form/YAML metadata editing
• Dynamic language variants in Editor (add language + optional AI translation)
• Variant auto-resolution supports both id styles: cv_<language>_<target> and cv_<language>_<iteration>_<target>.
• Language sync modal in Editor (pick source/target language pair with timestamp visibility)
• Keywords workspace (gap analysis and prioritization)
• Theme support (light/dark/system + template themes)
• Print Room photo customization modes (default/circle/square/original-ratio/off)
• Public fictional sample profile (John Doe)
• Tracked example company metadata plus an untracked personal company metadata file
• Privacy hardening for local/private artifacts (personal CVs, local DB/snapshots, editor/runtime files) to keep them out of git by default

Repository Layout

• apps/web/: Next.js web application
• services/parser/: FastAPI parser service scaffold
• packages/schemas/: shared schema/constants
• packages/render-core/: shared rendering primitives
• data/: sample CV and template mapping data
• templates/: template definitions and assets
• deploy/systemd/, deploy/nginx/: Linux deployment references

Documentation for Contributors and Agents

Process templates (adapted from shared defaults), checklists, and product specs:

• docs/DOCUMENTATION_INDEX.md — full index
• AGENTS.md — AI development operating contract
• CONTRIBUTING.md — human contributor quick start

Prerequisites

• Node.js >= 22
• npm >= 10
• Python 3.12.x (for parser service)

Quick Start (Local)

npm run bootstrap
npm run dev

Optional parser service (second terminal; scaffold only, see services/parser/README.md):

npm run dev:parser

When exposing the app beyond localhost, set MFCV_API_TOKEN in .env and send it as Authorization: Bearer … or x-mfcv-api-token on mutation/analysis routes.

Quality checks:

npm run lint
npm run typecheck

Keywords troubleshooting:

• If the Keywords tab shows empty JD results despite existing cache data, set SQLITE_BIN to your sqlite executable path (for example export SQLITE_BIN=/home/linuxbrew/.linuxbrew/bin/sqlite3) before starting the web app.

Production Build

npm run bootstrap
npm run build
npm run start

Default web runtime port is 3000 unless overridden by environment.

Hosting Guide

Windows 11

Recommended stack:

• app process: Node.js (npm run build && npm run start)
• optional parser process: Python uvicorn
• reverse proxy / TLS: Caddy (recommended) or IIS reverse proxy

Steps:

1. Install Node.js 22+ and Python 3.12.
2. Clone repo and run npm run bootstrap.
3. Build web: npm run build.
4. Run web as background service using NSSM or Windows Task Scheduler:
   • program: npm
   • args: run start
   • working dir: repo root
5. (Optional) run parser service with its own NSSM entry:
   • cd services/parser
   • python -m venv .venv
   • .venv\Scripts\pip install -r requirements.txt
   • .venv\Scripts\uvicorn main:app --host 127.0.0.1 --port 8001
6. Configure Caddy/IIS to proxy HTTPS traffic to 127.0.0.1:3000.

Linux (Ubuntu/Debian/RHEL)

Recommended stack:

• app process: systemd service for Next.js
• optional parser process: systemd service for FastAPI
• reverse proxy / TLS: nginx or Caddy

Steps:

1. Install Node.js 22+, npm 10+, Python 3.12.
2. Clone repo and run npm run bootstrap.
3. Build web: npm run build.
4. Create systemd unit for web process (npm run start).
5. (Optional) create parser venv and systemd unit for uvicorn on 127.0.0.1:8001.
6. Configure nginx/Caddy reverse proxy and HTTPS certs.
7. Enable auto-start:
   • sudo systemctl enable --now <web-service>
   • sudo systemctl enable --now <parser-service> (if used)

macOS

Recommended stack:

• app process: Node.js (npm run build && npm run start)
• optional parser process: Python uvicorn
• service manager: launchd (LaunchAgent/LaunchDaemon)
• reverse proxy / TLS: Caddy or nginx

Steps:

1. Install Node.js 22+ and Python 3.12 (Homebrew recommended).
2. Clone repo and run npm run bootstrap.
3. Build web: npm run build.
4. Create launchd plist for npm run start in repo root.
5. (Optional) create parser launchd plist for uvicorn on 127.0.0.1:8001.
6. Configure Caddy/nginx to expose HTTPS and proxy to 127.0.0.1:3000.

Runtime Data and Privacy

• Public repo includes only fictional sample CV data.
• Personal/local CV history and planning artifacts are intentionally untracked.
• Keep real CV content in local/private files outside version control.
• OpenRouter key is stored in local .env as OPENROUTER_API_KEY when saved via UI.

Release and Changelog

• Changelog is the source of truth for release notes:
  • CHANGELOG.md
• Changelog governance and writing style:
  • CHANGELOG_GUIDE.md
• API reference (including post-1.0.0 additions):
  • docs/API.md
• MCP wrapper guide:
  • MCP.md

Documentation Map

License

This project is licensed under the MIT License. See LICENSE. Template-specific license metadata lives under each template folder.