Alfa home pagelight logo
  1. Getting Started
  2. How It Works

​
How Alfa Works

Alfa is built as a connected workflow rather than a single text box. Brand context flows into topic discovery. Topics flow into article generation. Completed runs flow into review, export, and publishing.

​
The Operating Model

1

Build brand context once

Create a brand profile with website context, voice guidance, target audience, products, competitors, internal links, and writing rules.

2

Turn the brand into a backlog

Discover topics automatically or import them from CSV or JSON. Alfa saves topic sets so the research phase is reusable.

3

Generate long-form content

Launch the article pipeline from a topic or keyword. Alfa tracks the run and can retry stalled or failed work.

4

Review and ship

Edit the article, inspect SEO metadata, export in the right format, or publish to WordPress for the current brand.

​
1. Brand Context

The brand profile is the system memory Alfa depends on. It is where you define:

  • what the company does
  • who the content is for
  • which competitors matter
  • how the brand should sound
  • which words to prefer or avoid
  • where important internal links should point

Alfa can prefill large parts of this using website analysis, then refine the voice profile with article URL analysis.

​
2. Topic Discovery and Planning

Alfa’s planning layer is not just a keyword list. It creates saved topic sets that can include:

  • funnel stage
  • keyword priority signals
  • search intent
  • rationale for why the topic matters
  • saved topics that can be generated later

This is what separates the planning step from the writing step. You can research first, then decide which pieces are worth converting into full articles.

​
3. The 8-Agent Article Pipeline

The codebase currently maps article generation to eight agent stages:

1. Content Architect

Interprets the keyword, brand, and search context to shape the article strategy.

2. Outline Writer

Structures the article and defines the section hierarchy.

3. Deep Researcher

Collects sources, facts, and supporting material.

4. Writer

Produces the first complete draft.

5. Fact-Checker

Reviews claims against the available research.

6. Sub-Editor

Tightens the draft, improves flow, and cleans the prose.

7. Art Director

Attaches or prepares visual context such as hero or body image data when available.

8. SEO Optimizer

Finalizes search-facing metadata and output polish.

​
4. Review State and Recovery

Alfa stores runs and exposes recovery flows:

  • active pipelines can be resumed
  • stale runs can be retried
  • failed runs can be restarted
  • completed runs can be reopened from article management

This matters because long-form generation is not a fire-and-forget action. The product includes state tracking so content work survives interruptions.

​
5. Article Review

The review workspace is where the raw pipeline output becomes something you can ship. Alfa exposes:

  • article status controls: draft, in review, published
  • article text and structure
  • SEO title, slug, and meta description
  • schema and metadata views
  • copy and export actions

​
6. Export and Publishing

The current app supports several end states:

  • copy as Markdown
  • copy as HTML
  • copy as plain text
  • export Markdown
  • export MDX
  • export HTML
  • export JSON
  • publish to WordPress

WordPress is the current direct publishing path. Other destinations are represented in the settings surface and broader Alfa roadmap.

​
7. CLI Parity

The Alfa CLI mirrors the same model:

  1. authenticate
  2. create or select a brand
  3. discover or import topics
  4. generate an article
  5. watch status
  6. export the result

That means teams can move between the UI and scripted workflows without switching concepts.

Brand Profiles & Voice

Learn what goes into the brand layer.

Topic & SEO Workflows

See how Alfa plans and scores content opportunities.

Publishing & Export

Review the final delivery options.

CLI

Run the same system from your terminal.