SMF·LOWFRUITS
Can a prompt replace LowFruits?
SEO & marketing — keyword research, backlinks and rank tracking
Exhibit tracking slip
Verdict
The scoring rule is trivial: look at the ten results, count how many are forums, Q&A threads or low-authority blogs, and flag the keyword if enough of them are. What you cannot build is the input. Every one of those SERPs has to be fetched from Google, at volume, from the right country, and that is a paid data feed with a real bill attached. The clever part is free; the part that makes it work is not.
Exhibit A — The prompt
Received on31.07.2026Build a SERP weakness scanner: a tool that takes a list of keywords the user already has and tells them which ones are worth attacking because the current top results are weak.
Core loop. For each keyword, fetch the top 10 organic results from a commercial SERP API whose key sits in the environment (make the provider a single swappable module — the tool must not scrape Google directly, and must say so in the README). Classify every result against a weakness ruleset stored as editable YAML:
- user-generated platforms (reddit.com, quora.com, forums, Stack Exchange sites)
- listicle and aggregator patterns in the title ("best", "top 10", "vs")
- domains not in a user-supplied list of known-strong sites for that niche
- results whose page title does not contain the keyword or a close variant
Score each keyword by how many of the ten results trip a rule, and show the results table sorted by that score, with the raw SERP stored alongside so any score can be audited later. Store every fetch in SQLite, keyed by keyword plus country plus date, and never re-fetch a keyword already pulled that day — the API bill is the running cost of this tool and it must be visible: show total queries used and estimated spend on the dashboard.
Out of scope: generating keyword ideas, estimating search volume, backlink data, and anything resembling a crawl of the wider web. The tool scores the keywords the user brings; it does not find them.Opening prefills the prompt — press enter to run it.
Exhibit B — What you lose
- B.1 keyword discovery — you have to bring the list, it will not generate one
- B.2 search volume and difficulty figures
- B.3 the curated database of weak-site domains, built over years
- B.4 bulk runs across thousands of keywords without a data bill
Prior art
Exhibit C — Why people still pay: proprietary web index and data acquisition
Because the subscription is really a bulk purchase of SERP data with an interface on top. Running the same keyword set yourself through a commercial API costs about the same, and then you still have to maintain the classifier.
Questions
Can I import my LowFruits keyword lists?
Yes. LowFruits exports CSV, and the importer only needs a keyword column plus an optional country code, so a saved export drops straight in. The weakness scores do not transfer; they get recomputed against your own ruleset on the first run.
What does it actually cost to run?
The SERP API is the whole cost. Commercial providers sit around one to three dollars per thousand queries, so scanning 500 keywords is roughly a dollar. There is no monthly floor, which is the one genuine advantage over a subscription if you only scan occasionally.
Why not just scrape the results and skip the API bill?
Because it stops working. Google blocks automated queries quickly, and building around that means proxies, CAPTCHA solving and constant repair, which costs more in time than the API costs in money. The prompt deliberately rules it out.
What is the one thing this never gets right?
Knowing which domains are actually weak. LowFruits has years of accumulated judgement about which sites rank on brand rather than merit. Your YAML ruleset starts empty and only gets good after you have manually corrected it against a few hundred real SERPs.
Related tools
Receipt