Stop paying $130+ each per month for Semrush and Ahrefs, switch to the free and open source alternative (OpenSEO + DataForSEO)

Ahrefs charges $129 a month. Semrush charges $140. Then both meter what you can actually do through credit systems, so you pay full price and still run out of clicks mid month. If your SEO work comes in bursts, you are paying enterprise money for a tool that mostly sits idle. This guide sets up the alternative: an open source tool, hosted for free, where you pay for data by usage instead of by subscription. Our total bill after several days of heavy research: about fifty cents.

The credit systems are the real scam, and users say so. Ahrefs customers describe burning through a month of credits in days: "Keyword research is not fun anymore. It's Full of FEAR... Every damn filter seems to take 1 credit." Another: "ran out of credits 12 days into my billing cycle doing keyword research for a client deadline. Now I'm just... stuck?" You pay full price and still ration your usage.

The two pieces that work together

DataForSEO is where the data comes from. It is not a tool, and knowing that upfront saves confusion: there is no dashboard, no rank tracker interface, no reports. It is the wholesale data layer, an API selling keyword volumes, SERPs, backlinks and audits by the request, and a meaningful chunk of the SEO tool industry quietly builds on it. A SERP query costs about $0.0006. Fifty keyword lookups cost a fraction of a cent. The most advantageous part about this setup, is that YOU ONLY PAY FOR THE TOKENS THAT YOU BUY. No monthly fixed fee, you only pay for what you plan to use.

OpenSEO is the tool on top. MIT licensed, at github.com/every-app/open-seo, it gives DataForSEO the interface it lacks: projects, keyword research, rank tracking, site audits, competitor analysis, and a Google Search Console integration that pulls your real clicks and positions. It deploys to Cloudflare Workers and runs entirely on free tier services: Workers, D1, KV, R2. If it seems surprising that all of that hosting costs nothing, we wrote up how Cloudflare's free tier works and why separately. This app is a working example of it.

The setup

If you already have Claude Code setup with Cloudflare, then just prompt Claude to set this up for you, and it will. If you need to do the setup manually, then here is the guide:

Fork the repo, set two secrets, deploy:

pnpm install
npx wrangler secret put DATAFORSEO_API_KEY
npx wrangler secret put BETTER_AUTH_SECRET
pnpm run deploy

Write down BETTER_AUTH_SECRET somewhere safe. It encrypts stored OAuth tokens, and rotating it disconnects every linked account.

Three things that cost us time, so they do not cost you any:

1. Verify the DataForSEO account before assuming anything is broken. Until verification completes, the balance endpoint works while every data endpoint refuses, which looks exactly like a broken integration and is actually an unverified account.

2. Lock the app down, then confirm you did. It holds your API keys and search data. We put ours behind Cloudflare Access, and found that the Access toggle in the Workers dashboard did not actually create an Access application. Verify the protection exists rather than trusting the switch.

3. Enable the Search Console API in your Google Cloud project before connecting Search Console. Without it, every connection attempt reports as expired immediately after succeeding. The OAuth is fine. The API is off.

The part that changes the work

OpenSEO ships an MCP server, so Claude Code queries everything directly. SEO research stops being a dashboard you operate and becomes questions you ask:

That last one is the highest return activity in SEO, and it is tedious enough by hand that most people never do it. Asked as a question, it takes a minute. No credits were consumed. Nobody rationed anything.

What you give up, honestly

Two things. Ahrefs' backlink index is genuinely bigger: roughly 35 trillion links with history against DataForSEO's roughly 2.8 trillion live links. If deep backlink forensics is your daily work, that gap is real. And if you run continuous rank tracking across thousands of keywords, the per query math eventually favours a subscription; the crossover sits far above what a typical agency or in house team doing burst research will ever hit.

Worth knowing on the data quality question generally: the paid tools' numbers are estimates too. Users regularly find Ahrefs traffic figures off by large multiples against their own Search Console, and the classic thread is "SEMrush says volume of 20, Ahrefs says 700. Which to trust?" Switching is not trading truth for approximation. It is trading one estimate for another at a hundredth of the price, with your own Search Console data, which is the only real ground truth, wired in beside it.

In my opinion, you get about 90% of the performance, for a tiny fraction of the price