OpenSEO — a self-hostable SEO suite (keyword research, rank tracking, backlinks, site audits, competitor and AI-visibility reports) that talks to coding agents over MCP.
Alternative to: Semrush SEO at $139.95/mo billed monthly, or $117.33/mo billed annually, for 5 websites and 500 daily tracked keywords, plus $45/user/mo for a second seat · Ahrefs Lite at $129/mo for 1 user, 1,000 credits/mo and 750 tracked keywords · Moz Pro Starter at $49/mo for one website
Cost to run: free to self-host, but the data is not free. OpenSEO needs your own DataForSEO key, and DataForSEO's minimum top-up is $50, with no expiry and no monthly commitment. Google SERPs run $0.60 per 1,000 top-10 results on the standard queue, $2.00 per 1,000 in live mode; a top-100 pull is ten times that. Search volume via Google Ads standard queue is $0.06 per task of up to 1,000 keywords, which the vendor prices as $60 per million keywords. Backlink rows are $0.024 per request plus $0.000036 per row. The project also runs a hosted version at $10/month, which by its own README adds 28% on top of every DataForSEO request it makes for you.
Licence: MIT — free for commercial use.
Runs on: any machine that runs Docker, or Cloudflare's free plan for a shared internet-facing instance.
Requires: a DataForSEO API key. Docker for the simple path, a Cloudflare account for the recommended one. The README steers self-hosters toward Cloudflare over Railway, Coolify or Dokploy.
Traction: 17,816 stars · 2,256 forks · 522 commits onmain· open/closed issue counts unknown; issue and PR numbering has reached at least #303.
Liveness: very live — 82 commits in the last 30 days, 263 in the last 90, 19 authors in the last year, last commit 2026-09-03.
Maturity: v0.1.7, released 2026-09-02, across 36 tags. A sub-1.0 version number on a tool people are pointing at production sites.
Links: github.com/every-app/open-seo · openseo.so · no substantial Hacker News thread exists; three submissions across 2026 scored 3, 2 and 1 point, with 0 comments between them.
Figures as of 2026-09-08.
What it does
OpenSEO runs the six workflows that make up most of what people actually open Semrush for: keyword research, rank tracking, competitor insights, backlinks, site audits, and AI visibility. It ships as a web app you host yourself, and it does not carry its own data. You bring a DataForSEO key and pay that vendor directly for every lookup.
The second half is the part that has no obvious paid equivalent. OpenSEO exposes an MCP server, so an agent like Claude Code can query your SEO data directly, and it ships prebuilt Agent Skills that walk an agent through common tasks. The install is npx skills add every-app/open-seo.
What it replaces, and what that costs
The three incumbents, from each vendor's own pricing page:
Semrush — the tier now sold as SEO: $139.95/mo billed monthly, $117.33/mo billed annually. Covers 5 websites and 500 keywords tracked daily. A second user is $45/mo on top.
Ahrefs — Lite: $129/mo, one user, 1,000 credits a month, 750 tracked keywords, 6 months of historical data. A second user is $40/mo, up to two.
Moz Pro — Starter: $49/mo, described by Moz as "designed specifically for individuals managing a single website."
Against that, OpenSEO's own costs are metered rather than fixed. There is no seat price and no monthly floor, only DataForSEO's $50 minimum top-up, which does not expire and is only drawn down when a request returns a result.
That difference is the whole pitch, and it is also where the comparison has to stop. These numbers do not subtract. Semrush and Ahrefs sell a subscription that includes their own crawled index; OpenSEO sells you nothing and routes you to a third party's data. Whether $50 of DataForSEO credit lasts a month or a morning depends entirely on how many SERPs you pull and how deep. A person tracking 500 keywords daily and a person running one audit a week are not on the same bill, and neither one is on Semrush's bill.
Worth flagging for anyone pricing the hosted option: $10/month buys you the app, not the data. The DataForSEO charges still land, marked up 28%, which the project states plainly in its own README rather than burying.
Is it still alive?
Yes, and busily. 82 commits in the last 30 days and 263 in the last 90, from 19 authors over the past year, with the most recent landing 2026-09-03. Thirty-six tags, latest stable v0.1.7 on 2026-09-02, whose release notes describe on-demand SERP depth and fixes to Search Console reporting through MCP.
The maintainer concentration is real and worth stating: Ben Senescu accounts for 486 of the recorded contributions across two email addresses. The next contributor has eight. This is one person's project with an active crowd around it, not a distributed team.
The issue tracker answers. Issues opened on 2026-09-03 and 2026-09-04 are numbered in the 270s and 280s against a project that has reached at least #303, which is a tracker moving in days rather than months.
What people say
Hacker News never picked this up. Three submissions exist and none drew a single comment: a Show HN on 2026-04-27 titled "Show HN: Open-source Semrush alternative just passed 1.6k stars" at 2 points; "OpenSEO" on 2026-07-21 at 3 points; and "OpenSEO – open-source SEO Platform" on 2026-08-29 at 1 point. That is the finding, not a gap in the research. A repo that went from 1,600 stars in April to 17,816 in September did it without a Hacker News thread. Reddit is blocked from the research sandbox, so no vote counts from it are claimed here either.
The tracker is where the conversation is, and one theme dominates. Citing titles rather than full threads: lucafaggianelli opened #281, "Support for alternative analytics like Plausible," on 2026-09-03; patente-ar opened #287, "Connect Matomo as an analytics source alongside Google Analytics," on 2026-09-04; CodeAndWeb opened #286, "Add read-only Cloudflare Analytics integration," on 2026-09-04. Three people, two days, all asking the same thing: let me use this without Google.
The other signal is outside contributors arriving with work already done. richardross23 opened #274 and #272 on 2026-09-03, both titled as proposals with patches attached, on crawl budget being spent on images and tracking-parameter duplicates, and on Shopify storefronts returning 429 to the audit crawler.
Dissent, such as it is, is mundane: jhmonroe opened #271 on 2026-09-02, titled "speed up build time?"
The catch
The version number is v0.1.7. Thirty-six tags in, the project has not called itself 1.0, and the entry takes that at face value rather than arguing with it.
The data question is the real one. DataForSEO is a competent SERP and keyword vendor, but its backlink index is its own, not Ahrefs'. Anyone switching for backlink research is changing data sources, not just interfaces, and the honest test is to run a domain through both before cancelling anything.
Self-hosting has a recommended path and a discouraged one. The README points at Cloudflare and explicitly says that unless you already self-host confidently, Railway, Coolify and Dokploy are not yet the easy route, with better support promised "in the next few months." Docker is offered for testing on your own machine rather than for a team.
Open and closed issue counts are not obtainable from this research sandbox, so the traction row reports the highest number reached instead of a ratio.
My read
The interesting thing here is not that someone built a cheaper Semrush. It is that the pricing model changed shape. Semrush and Ahrefs sell you a ceiling you rent whether you touch it or not, and the seat fees are where they really get you — $45 a month for a second Semrush user is most of a Moz subscription for one extra person's login.
OpenSEO throws that out and charges you for the queries you actually run. For anyone who opens an SEO tool twice a month, that is a completely different economic proposition.
But the line this entry has to hold is that the data is not the same data. The workflows subtract. The backlink index does not. DataForSEO is not running Ahrefs' crawler, and anybody who tells you a $50 top-up replaces a $129 subscription is comparing an interface to an index. Run a domain through both, look at what comes back, and then decide. That is the only version of this comparison that survives contact with a real site.
Credits
Ben Senescu (bensenescu, X: @bensenescu) maintains OpenSEO and wrote 486 of its recorded contributions. Jeremy Rivera (jeremypenguin-max) landed the most recent commit on main. Issue reporters quoted or cited by title: lucafaggianelli, patente-ar, CodeAndWeb, richardross23, jhmonroe.
Figures came from the OpenSEO repository and its README, ungh.cc for counts, a blobless clone for commit history, and the published pricing pages of Semrush, Ahrefs, Moz and DataForSEO. The Hacker News submission scores came from the Algolia HN API.