Description
SEO used to be about ranking in the top ten; now people ask an AI directly and your brand simply is not in the answer. That is the problem GEO addresses — and in practice it falls apart operationally: source material sits in cloud storage, drafting happens in documents, review happens in chat, publishing is manual, and the results scatter across several dashboards. Nobody can say which article an AI actually cited, or why.
GEOFlow puts the whole chain in one admin panel: trusted knowledge → AI generation → quality gate → human sign-off → multi-site distribution → results flowing back, with the source, evidence, review decision and publication state recorded at each step so a post-mortem can trace all the way back.
Aimed at corporate sites, GEO sub-channels and industry reference sites. Self-hosted on PHP, Laravel and PostgreSQL, open-source under AGPL-3.0, and wired to OpenAI-compatible endpoints so swapping models needs no code changes.
Start work from one sentence: rather than hunting through menus, say something like "analyse brand visibility over the last 30 days, find content whose citations dropped, and create optimisation tasks" — it reaches into the analytics, the visibility diagnostics and task orchestration and dispatches the work from one conversation.
A quality gate before publishing: AI-written drafts pass an automated inspection first, and anything that fails never reaches a human reviewer. The verdict is stored with its supporting evidence rather than reduced to a score.
Human sign-off is not skipped: passing the machine check still requires a person to approve. A review centre, trash and bulk operations are all there, and who approved what and when is recorded for team work.
AI visibility tracking: a dedicated area watches whether your content is being cited by AI, reported as brand visibility rate and citation movement — data a conventional SEO dashboard does not give you.
Multi-site distribution: one knowledge base feeds several sites, publishing separately to the corporate site, sub-channels and hosted sites, with a browser extension assisting manual posting to third-party channels.
Knowledge base with evidence snippets: imported material is split into retrievable evidence snippets for RAG, so generated content traces back to a specific source instead of the model improvising.
Deployment built for production: blue-green deployment and automatic migration are provided, updates are signed, and the project advises running a tagged Release in production rather than following the main branch.
GEOFlow puts the whole chain in one admin panel: trusted knowledge → AI generation → quality gate → human sign-off → multi-site distribution → results flowing back, with the source, evidence, review decision and publication state recorded at each step so a post-mortem can trace all the way back.
Aimed at corporate sites, GEO sub-channels and industry reference sites. Self-hosted on PHP, Laravel and PostgreSQL, open-source under AGPL-3.0, and wired to OpenAI-compatible endpoints so swapping models needs no code changes.
Features
Start work from one sentence: rather than hunting through menus, say something like "analyse brand visibility over the last 30 days, find content whose citations dropped, and create optimisation tasks" — it reaches into the analytics, the visibility diagnostics and task orchestration and dispatches the work from one conversation.
A quality gate before publishing: AI-written drafts pass an automated inspection first, and anything that fails never reaches a human reviewer. The verdict is stored with its supporting evidence rather than reduced to a score.
Human sign-off is not skipped: passing the machine check still requires a person to approve. A review centre, trash and bulk operations are all there, and who approved what and when is recorded for team work.
AI visibility tracking: a dedicated area watches whether your content is being cited by AI, reported as brand visibility rate and citation movement — data a conventional SEO dashboard does not give you.
Multi-site distribution: one knowledge base feeds several sites, publishing separately to the corporate site, sub-channels and hosted sites, with a browser extension assisting manual posting to third-party channels.
Knowledge base with evidence snippets: imported material is split into retrievable evidence snippets for RAG, so generated content traces back to a specific source instead of the model improvising.
Deployment built for production: blue-green deployment and automatic migration are provided, updates are signed, and the project advises running a tagged Release in production rather than following the main branch.
