Source of record
Publish the canonical asset first
Commercial pages, cases, and owned research go live on humanswith.ai with metadata, schema, internal links, and a stable URL.
Module 04 · Content Publisher
The publishing agent for AI Search work. Approved packets become canonical assets with proof.
Content Publisher turns approved work into live, traceable, canonical content.
In 2026 we use it as the operator-controlled handoff between ContentOS and the live website. ContentOS produces the packet. Website Agentic Optimization keeps the site readable. Content Publisher chooses the route, prepares search metadata, hands off schema, builds distribution drafts, and records proof for the next visibility scan.
Packet
Route
Proof
Publish receipt
- Route
- Metadata
- Schema
- Proof
Published for
- ChatGPT
- Claude
- Perplexity
- Gemini
- Grok
- DeepSeek
- Kimi
- Google AIO
- Copilot
Why publishing needs an agent
Most teams lose the loop after the draft is approved.
AEO/GEO work is not finished when the article reads well. A draft can pass editorial review and still never become a source AI systems can cite.
Route
The asset needs a canonical home
The publisher chooses the stable URL, checks route uniqueness, and keeps the page tied to the visibility gap that triggered the work.
Structure
The page must be readable by machines
Metadata, schema, internal links, hreflang, and OG data are prepared before the rebuild instead of patched after publication.
Proof
The run ends with evidence
Distribution drafts, deploy receipts, sitemap notes, and the next Hermes or AI Visibility scan stay connected to the same packet.
The job is not "push publish."
The job is to make the asset reachable, canonical, structured, linked, distributed, and measurable in the next Hermes scan.
01 · Approved packet
Receive the ContentOS work
The agent starts from a governed packet. It receives final copy, source constraints, quality report, schema handoff, visual needs, and explicit operator approval.
02 · Publish safely
Route, metadata, schema, rebuild
It checks slug uniqueness, canonical intent, title and description, OG data, structured-data handoff, links, hreflang needs, and rebuild readiness.
03 · Measure again
Distribution and proof receipts
After publication, it records the route, channel drafts, sitemap notes, deploy receipt, and the next Hermes or AI Visibility scan task.
Canonical-first distribution
The website is the source of record. External channels are adaptations.
Adaptation
Adapt for high-fit channels
English distribution can use LinkedIn, Medium, dev.to, partner blogs, and newsletters. Russian distribution can use VC.ru, Dzen, Telegram, and partner media.
Attribution
Link back to the source
External posts point back to the canonical page. The link and source framing matter more than assuming platforms honor rel=canonical.
Proof
Re-scan after crawl time
Hermes or AI Visibility checks whether the new asset changes mentions, citations, source context, or recommendation language.
Route and metadata
Every asset has a job
The publisher ties each route to a visibility gap, buyer intent, language, canonical URL, OG image, and priority link target.
Preflight
No duplicate route surprises
Duplicate slug checks, block checks, schema handoff, link gates, hreflang checks, and rebuild preflight keep the site from regressing.
Proof
Publication becomes measurable
The output is not only a live URL. It is a receipt: what shipped, where it links, which channel adaptations are ready, and what scan should prove the next result.
Where it fits in the workspace
Content Publisher is the bridge between production and proof.
Before · Produce
Workspace ContentOS
Creates the brief, final draft, schema handoff, quality report, and approval packet.
Open ContentOS →During · Publish
Content Publisher
Turns the approved packet into a canonical route, channel drafts, rebuild proof, and follow-up scan task.
Foundation · Optimize
Website Agentic Optimization
Makes sure the site layer is readable, structured, crawlable, and safe for the asset to compound.
Open Website Agentic Optimization →FAQ
Questions teams ask before publishing through an agent.
What does Content Publisher do?
Content Publisher turns an approved ContentOS packet into a controlled publishing run. It handles route selection, metadata, schema handoff, canonical checks, distribution drafts, rebuild preflight, and post-publish proof. It is the content manager agent in the Humanswith.ai workspace.
Does it publish without human approval?
No. The default policy is approval-first. ContentOS prepares the packet, humans approve sources and claims, and Content Publisher executes only within the approved route, channel, metadata, and distribution scope.
How is this different from a CMS scheduler?
A scheduler moves a draft from queued to live. Content Publisher checks the canonical page choice, route uniqueness, metadata, schema, internal links, hreflang, and deploy traceability before the asset ships.
Which channels does it support?
The product pattern is canonical-first: publish the commercial page, article, or case on humanswith.ai first. Then adapt it for LinkedIn, Medium, dev.to, newsletters, VC.ru, Dzen, Telegram, or partner sites depending on language and audience.
How does it help AEO/GEO?
AI engines cite assets that are reachable, canonical, structured, and linked into the entity graph. Content Publisher keeps the route, metadata, schema, sitemap, and distribution receipts connected to the visibility gap that triggered the work.
What proof does a publishing run leave?
A publishing run should leave a route, canonical URL, metadata snapshot, structured-data status, link-gate status, and rebuild or deploy receipt. When applicable, it also records sitemap or re-indexing notes and a follow-up scan task for Hermes or AI Visibility.
Is this only for blog articles?
No. The same workflow applies to commercial pages, case studies, glossary pages, research adaptations, comparison pages, social distribution drafts, and canonical updates to existing pages.
Turn one approved packet into a measurable publication run.
We show the full loop: visibility gap, ContentOS packet, canonical page, distribution drafts, and proof scan.
You see how a marketing engineer operates the workspace instead of hiring separate freelancers and agencies for copy, content management, SEO, website changes, and reporting.