Automotive WatchDog is a serverless news-intelligence pipeline: it pulls automotive-industry RSS feeds, uses AI to filter and categorize each story, and publishes a live map plus a searchable historical archive. Built incrementally, each phase shipped and verified live before the next started.
Roadmap
Live news map
ShippedDaily pipeline: fetch, AI-enrich, rank and publish the day's top automotive stories to an interactive map.
Historical archive
ShippedEvery enriched story is permanently archived (not just the day's top picks), since the operational database only keeps 14 days.
Date Tracker
ShippedBrowse the archive by date range: presets or custom, period-over-period comparison, trend breakdowns, and a historical map.
AI Briefing
ShippedA daily AI-written executive summary for each standard period (last 7/30/90 days, year to date) — read it on the Date Tracker page.
Company profiles
Shipped (12/12 companies)Evidence-backed company priorities and strategy, sourced from official documents and linked to page citations. All 12 companies are live: Toyota, Volkswagen, Tata Motors, Ford, Honda, Mahindra, BYD, BMW, Maruti Suzuki, General Motors, Stellantis and Mercedes-Benz (the latter three from their 2024 report, one year behind the rest, pending a reachable 2025 source).
Strategy comparison
ShippedSee how all 12 companies' stated priorities compare across seven themes — electrification, software, manufacturing, geography, capital allocation, sustainability, and safety — each still linked back to its source document and page. Read it as a tab on the Company Profiles page.
Strategy vs. signal
PlannedConnect declared company strategy to real-time news: flag supporting, contradicting, or emerging evidence.
Infrastructure cost
Entirely serverless on AWS: Lambda, DynamoDB, SQS, S3 and CloudFront, scaling to zero between runs. Based on real AWS Cost Explorer token/request usage data (not a guess), priced at Amazon Nova's and Amazon Location Service's published on-demand rates, at the pipeline's actual current volume — re-derived Sep 9, 2026 with two more weeks of real usage, including Company Profiles and live map-tile traffic for the first time.
| Everything except AI (Lambda, DynamoDB, SQS, S3, CloudWatch) | $0.00, inside AWS free tier |
| Story enrichment (Amazon Bedrock, Nova Lite) | ~$0.50 / month |
| Trend extraction + AI Briefing (Amazon Bedrock, Nova Pro) | ~$0.51 / month |
| Company Profiles extraction (Amazon Bedrock, Nova Pro) | ~$0.04 / month steady-state |
| Map basemap (Amazon Location Service tiles) | ~$0.21 / month (~5,300 tiles) |
| Total, averaged | ~$1.27 / month |
Story enrichment and trend extraction/Briefing are re-derived from 14 days of real Nova token usage; the Nova Pro figure actually came down slightly from the prior estimate, which had projected AI Briefing's steady-state cost from just its first live day. Company Profiles' one-time cost to onboard all 12 companies (backfilling every document from scratch) was about $0.52 total, already spent — the figure above is the much smaller ongoing cost of re-processing a company's document when it publishes a new one (roughly once a year per company). The map basemap is now a real, measured number (real tile-request volume × its published $0.04/1,000 rate, confirmed against AWS's own price list) rather than an unestimated usage-based line. Excludes the shared CloudFront distribution's own baseline traffic costs for pranavjanardhanan.com, which exist independently of this app. Runs on Amazon Nova rather than Claude models due to a temporary Bedrock access limitation, so cost would run somewhat higher on Claude, still in the low single-digit dollars per month at this volume.