Open source SEO tools, in one self-hosted content engine
AltoRank is an open-source alternative to Outrank and Jasper: research a keyword, generate the brief, write the article, score it, and publish it to any of 10 CMS platforms. Open source, with the whole product in the open: agency multi-tenant, white-label reporting, and the editorial approval gate included, not held back for a paid tier. Self-host it for nothing, or let us run it.
No ee/ directory, no held-back tier
"Open core" usually means the interesting parts are the paid parts. Here the features people actually buy a tool for are in the free build. What you pay for is hosting, the model bill, and not being the one who maintains it.
| Capability | Typically | In the free build |
|---|---|---|
| Multi-client workspaces | Reserved for an agency or enterprise tier | In the free self-hosted build |
| White-label reports | Premium add-on, or locked to the top tier | In the free self-hosted build |
| All 10 CMS integrations | Metered by connection count | In the free self-hosted build |
| Editorial approval workflow | Often absent entirely, autopilot by default | In the free self-hosted build |
| The SEO scoring algorithm | A black-box number you cannot inspect | Readable source you can audit and change |
Run it yourself. Put your brand on it. Keep the margin.
The managed-service providers moving into AI search will happily sell your client an outcome directly, and cut you out of the relationship while they do it. The open-source build is the other side of that trade: the same execution layer, running on your infrastructure, under your brand.
White-label reporting and multi-client workspaces are in the free build specifically so anyone can stand this up without a licence conversation. You own the client relationship, you set the retainer, and the software cost is your infrastructure bill rather than a per-seat line item that scales against you.
Self-hosting it for your own clients carries no publication obligation, see the licence question below.
Ask three AI-visibility tools the same question, get three answers
This is the standing complaint about the entire AI-visibility category, and it is a fair one. Point several tools at the same brand with the same prompts and the citation rates disagree. Because every vendor's methodology is private, there is no way to tell which one is closer to the truth. You are asked to trust a number whose derivation you cannot see.
Open source is the only real answer to that. When the prompt set, the sampling method, and the scoring algorithm are readable, a disagreement between two tools becomes a question you can settle by reading code instead of a matter of vendor faith. Ours are in the repository. Read them, disagree with them, and open an issue if the methodology is wrong.
How we track AI citations for clients → : the frozen prompt set, logged-out sampling, and reporting cadence, written out in full.
Free to self-host, priced when we run it
Self-host
The whole engine, open source. Your infrastructure, your API keys, no feature gates.
Managed
No API keys to manage, because model and data costs are included. For solo operators and agencies running one or two brands.
Agency
For agencies running content across a full client roster. Everything metered on output, not seats or workspaces.
Full breakdown on the pricing page.
The whole engine, in the open
The open release is the whole product rather than a first slice: research a keyword, build a brief, write the article, score it against the on-page checks, lint it for the filler phrasing that marks generated text, and publish it to your CMS, plus the dashboard, agency multi-tenant and white-label reporting. It also runs from an AI assistant over MCP, point Claude Code at it and it can research and draft against your own brief. A packaged command-line tool is not built yet.
The minimum to do useful work is a single Anthropic key. Keyword volume and SERP data need a DataForSEO key on top, and without one the brief step degrades to prompt-only rather than failing. Publishing defaults to draft status, always.
The hosted dashboard, agency reporting, and billing come after. Saying so up front is more useful than a roadmap that implies everything lands at once.
Common questions about the licence
Is AltoRank really open source?
Yes. AltoRank is open source, and the whole product is in the public repository: agency multi-tenant, white-label reporting, the approval workflow, and all 10 CMS adapters. There is no separate "enterprise" directory and no paid tier held back from the source.
Can I run this for my own clients, commercially?
Yes, and for free. Self-host it, run it as-is for your clients, and you owe nothing and publish nothing: commercial use on behalf of clients is exactly what the licence is for. One obligation exists and it only reaches people rebuilding the product, not people using it. If you modify the source and offer that modified version to others over a network, your changes have to be shared back on the same terms. Running an agency on it is free and unencumbered. Relaunching a modified copy as a competing service is not.
What stops a competitor taking it and reselling it?
Nothing stops anyone self-hosting or forking it, and that is deliberate. What the licence does stop is a competitor taking the multi-client workspaces, the white-label reporting and the exchange, changing them, and running the result as a closed commercial service without sharing those changes back. That fence is the reason the whole product can ship in the open instead of holding features back in a paid tier: what is defensible here is software, not a metered API.
If I can self-host it for free, what am I paying for?
Hosting, the model bill, and not maintaining it. The self-hosted build is the same software, but you supply the infrastructure and your own API keys, and you carry the upgrades. The paid tiers exist because someone else runs it, the model and data costs are included, and the output limits are metered. Nothing is withheld from the free build to make the paid one attractive.
Can I white-label AltoRank for my clients?
Yes, and it is in the free self-hosted build rather than gated behind a premium tier. That is the point of the model for agencies: run it on your own infrastructure, put your own brand on the reports, and keep both the client relationship and the margin. Your clients never need to know what is underneath.
What ships first?
It already has. The repository is public and it is the whole product, not a stripped first slice: the engine, the dashboard, agency multi-tenant, white-label reporting, the approval gate and all 10 CMS adapters. An MCP server is included, so an AI assistant can research and draft against your own brief. The one thing that is NOT built yet is a packaged command-line tool; the plan was to lead with one, and shipping everything at once turned out to be the smaller step.
Does the approval gate survive in the open-source build?
Yes. The gate lives in the publish step itself, so scheduled jobs and bulk actions cannot route around it, and it has two ways through: a person approves the draft, or a person turns on automatic publishing for the workspace, which runs the same sourcing and audit checks after a hold they can veto. Both are recorded under a name; no agent can write either. Every existing workspace stays on manual review unless someone changes it.
This page describes how the licence is intended to work and is not legal advice. The full licence text ships with the source, and if the boundary matters to your business, take your own advice.
Be the site the assistant names.
Add a domain and the first draft is written while you watch. Every publish is your decision, a click on the draft or a rule you set and can hold, and you can read the source or self-host it free.
Add a domain, it sets up your workspace