SEO Keyword Research Tool Setup

We build websites for restaurants and need a tool that turns a single restaurant into a complete keyword plan our dev team builds from. Given a restaurant, the tool pulls its Google Business Profile and researches the best keywords across three dimensions — its own city, surrounding towns it can realistically rank in, and its cuisine/signature dishes — each with real monthly search volumes — then outputs a page-by-page plan (target keyword, volume, category, URL slug) as JSON/CSV. It also crawls the restaurant’s existing website to flag new vs. updated pages and on-page SEO defects. We want to scrape the google reviews as well to pull out the best dishes and services offered by the restaurant. We have a working prototype, so the problem is well-defined. We want it rebuilt properly by someone who knows local SEO and AI development better than we do. Must have: proven local-SEO expertise · strong TypeScript/Node or Python · third-party API integration (DataForSEO or similar keyword/SERP API) · hands-on LLM integration, with sound judgment on where AI fits vs. deterministic rules. Full technical scope is attached. To apply, please answer: 1. Link or describe an SEO/keyword tool you’ve built programmatically. 2. How would you stop the tool recommending irrelevant keywords (e.g. “sushi” for a ramen shop)? 3. Where would you use an LLM vs. fixed rules? 4. Have you used DataForSEO or a comparable API, and how do you keep per-run cost predictable?