feat: corporate ATS scrapers (Ashby, Greenhouse), O&G, government boards, posted_date fixes

add WorkingNomads + SkipTheDrive scrapers (social media, CS, entry-level focus)

fix: soup.find(name=...) keyword clash in extract-url

fix: move import hashlib to top, wrap extract-url in try/except

feat: major quality & diversity overhaul

feat: add creative HR email generation on job-detail page

fix: api key persistence — delete old keys before insert, add logging, verify saved keys in response

fix: 'DEPLOY IN 3 MINUTES' → 'APPLY IN 3 MINUTES'

fix: landing page polish — Register nav, 87 visitor counter, slower scroll-reveal animations, redesigned job cards with match quality rings, better last-scraped display

feat: retro-futuristic neon landing page redesign with 3D canvas, job carousel, testimonials carousel

fix: larger CV fonts for readability, 4-7 bullet points per role, force experience extraction from raw text, ban summary text in education

feat: incrementing counter file naming (01_name_company) + fix API key overwrite bug + better LLM logging

fix: RGBColor is a tuple subclass, use indexing not .red/.green/.blue attrs

fix: sequential save in convertPlainText to avoid race condition on raw_text

fix: preserve raw_text in save_cv — don't overwrite with empty on structured CV saves

feat: pass raw CV text to LLM during tailoring for full context

fix: save plain text CV on convert + load raw text on cv-editor page load

fix: isolate ALTER TABLE in own transaction to prevent PG InFailedSqlTransaction + auto-save CV editor

fix: move ParseCVRequest before parse_cv endpoint to resolve NameError

fix: restore missing generate_cv_pdf function definition

feat: Impeccable UI — OKLCH colors, tinted neutrals, solid buttons, cleaner cards, better type scale, LLM CV parser

feat: CV diversity profiles, Gemma provider, transferable skills, paste persistence, improved fallback

fix: cast date_found to TIMESTAMP in PG comparisons + migration

fix: _exec must return cursor for both PG and SQLite

fix: _exec returns conn for SQLite instead of cursor + add 6 new job boards (Bruntwork, ScaleArmy, crypto/web3)

fix: add error logging to /api/stats endpoint

fix: psycopg2 needs cursor.execute not conn.execute

feat: add Neon PostgreSQL support via DATABASE_URL + Fly.io deployment config

fix: remove broken docx_generator and database late imports (same pattern as config bug)

fix: /api/config 500 from broken late import of config module

fix: scrapers, cron, design, filters, categories, graduate detection

Fix test endpoint structure for ngo/intl/hi boards

Add test endpoints for ngo/intl/highimpact scrapers

Fix highimpact.py syntax and method structure

Fix method name mismatches for 80000hours, himalayas, remoteworkng, and realworkfromanywhere

Fix brotli compression issue in base scraper headers

Add raw HTML debug to test endpoint

Add debug logging to scrape_test endpoint

Fix scraper CSS selectors and add error logging

Add README.md with HF Space metadata

Add super admin user role with view toggle

Fix: dont redirect on login 401 so error message shows

Add password visibility toggle & password reset flow

Run scrape in background thread to avoid timeout

Improve mobile responsive layout

Add about.html with dev profile + how-to section

Fix API base URL for HF Spaces

Initial commit: Joblin platform