AEOForgedAEOForged
Join

Workflow Recipes

Copy-paste agent workflow patterns. Prefer consolidated verbs (create, improve, diagnose). Each recipe shows the tool chain, what to pass between steps, and what it costs.

Last updated

AEOForged workflow recipes with their tool chain length, credit cost, and primary use case.
#WorkflowToolsCostBest for
1Agent Writes (Recommended)5~10+ credits (research; score/factcheck free; improve 2–8 by mode)IDE / agent-authored articles
2Outline Only2~13 credits (10 research + 3 outline)Planning before writing
3Hosted Draft (create)120 credits (single/gate); draft-mode pipeline alias 12crDraft in AEOForged to edit
4Site AI-Ready (diagnose)4Diagnose/crawlability free; per-page research/improve as needed; llms.txt 3crSite-wide AI readiness
5Fix Programme Loop4Free for queue/verify/document; research/improve bill if usedClosing audit findings
6Original Research Study1Design 8cr + run (lane-priced) + draft 8cr; register freeCitable primary research
7Audit and Improve Existing Content3~5–16 credits depending on compare + refreshFixing an existing page
8Build a Content Strategy45 + (10–15 per pillar), or ~44 for create strategy modePlanning a content calendar
9Citation & Visibility Monitoring3Track free; visibility/SOV 10cr when you run the proxy toolTracking published content
10Composite Workflows (Legacy Aliases)40–44 credits (quality_gate free; strategy ~44; audit_improve ≤13; health 5/URL)One-call automation (legacy)
1

Agent Writes (Recommended)

The recommended path when you draft in Cursor or Claude Code: research first, optional outline, you write with research-backed citations, then score → improve → factcheck → save → client review gate.

Tool Chain

aeo_researchaeo_outline(optional)aeo_scoreaeo_improveaeo_factcheck

Cost

~10+ credits (research; score/factcheck free; improve 2–8 by mode)

How It Works

  1. 1REQUIRED: research the topic (pass project_id when a workspace article exists) — writing and submit-for-review are gated on research-backed sources
  2. 2Optional: outline for structure, then write markdown in the IDE using only research URLs as inline citations
  3. 3Score (free) — pass research; check writingQuality; optionally one improve(humanize) if needed
  4. 4improve(to-score) at most 3 times (if still below 80, stop and report weakest dimensions), then factcheck statistics (free)
  5. 5Save article → submit-for-review → poll gate-status at most twice; if wait, stop and tell the human to prompt you after Ready
  6. 6Ship live per work_access, then schema / optional track as needed

Agent Prompt Example

“Research 'answer engine optimization', I'll write the guide in Cursor using only those sources. Score and improve until ≥80, factcheck stats, then submit for review.”

2

Outline Only

Research + outline, then hand structure back to a human or agent writer — no hosted draft spend.

Tool Chain

aeo_researchaeo_outline

Cost

~13 credits (10 research + 3 outline)

How It Works

  1. 1Research with brief + project_id if the article shell exists
  2. 2Generate outline from research (contentType: article | faq | product | guide)
  3. 3Human or agent writes from the outline; continue with the Agent Writes score loop

Agent Prompt Example

“Research 'container security for DevOps' and produce an outline only — I'll write the article myself.”

3

Hosted Draft (create)

One-call hosted production via aeo_create (or aeo_full_pipeline alias): research → outline → write → polish with grounding. Prefer create over chaining granular tools when you want a dashboard draft.

Tool Chain

aeo_create

Cost

20 credits (single/gate); draft-mode pipeline alias 12cr

How It Works

  1. 1Optional: build/reuse client profile so voice and company facts flow into the draft
  2. 2Call aeo_create (or POST /api/v1/create) with topic + brief — async research steps may return 202 + poll_url
  3. 3Score in the dashboard; iterate with aeo_improve; factcheck before review
  4. 4Human edits; agent persists follow-ups with PUT .../article

Agent Prompt Example

“Create a hosted draft guide on 'answer engine optimization' for indie hackers — grounded research, then I'll edit in the dashboard.”

4

Site AI-Ready (diagnose)

Make the whole site readable and fixable for answer engines: diagnose (site-wide audit), crawlability, then per-page research → improve → score. Prefer aeo_diagnose over chaining complete_audit + workflow_health.

Tool Chain

aeo_diagnoseaeo_crawlabilityaeo_improveaeo_package

Cost

Diagnose/crawlability free; per-page research/improve as needed; llms.txt 3cr

How It Works

  1. 1aeo_diagnose (or complete_audit) with workspace_id — async by default → poll tool-runs
  2. 2aeo_crawlability on production — catch robots / llms.txt / JS blockers first
  3. 3Prioritise technical + page findings; per page: research → improve(aeo) → rewrite in repo → score
  4. 4package for JSON-LD / llms.txt when needed; optional agent-ready / agent-policy for operability

Agent Prompt Example

“Diagnose https://example.com for AI readiness, check crawlability, then improve the weakest money pages until they score ≥80.”

5

Fix Programme Loop

Work an audit action queue page-first: claim → apply fix → done → verify-page (system promotes verified) → document_fix. Craft over score on live marketing pages — never invent citations or set verified yourself.

Tool Chain

aeo_audit_queueaeo_action_updateaeo_verify_pageaeo_document_fix

Cost

Free for queue/verify/document; research/improve bill if used

How It Works

  1. 1Bootstrap handoff or onboard — read audit_engagement (snapshot, queue, open items) and work_access
  2. 2Claim: PATCH action-item → in_progress (MCP: aeo_action_update)
  3. 3Apply the fix (repo / CMS / PR per work_access); optional preview-fix first
  4. 4Mark done → document_fix (problem + fix) → verify-page { url, snapshot_id }
  5. 5Verified is system-only — if checks fail, reopen in_progress and fix again

Agent Prompt Example

“Work the open audit queue for this workspace: claim the highest-impact item, apply the fix in the repo, document it, and verify-page against the snapshot.”

6

Original Research Study

Produce measured original research the client can publish: design → run (deterministic instrument) → draft narrative from the frozen rollup → human publishes article/methodology/dataset → register as primary sources. High-level lane only — no invented statistics.

Tool Chain

aeo_study

Cost

Design 8cr + run (lane-priced) + draft 8cr; register free

How It Works

  1. 1design — propose question framing; validator enforces study rules (thin samples refuse later)
  2. 2run — async measurement only (no LLM inventing stats); poll tool-runs; depth floor must pass before draft
  3. 3draft — narrative sections quote the frozen rollup; methodology is code-authored
  4. 4Human publishes the three pages on the client domain → register (all-or-nothing live verification)

Agent Prompt Example

“Design and run a benchmark study on our category competitors, draft the article from the measured rollup, then register once we've published all three URLs.”

7

Audit and Improve Existing Content

Score a live URL, optionally compare a competitor, then refresh or improve(aeo/freshness). Prefer aeo_improve and aeo_diagnose composites when the task fits.

Tool Chain

aeo_auditaeo_compareaeo_improve(freshness)

Cost

~5–16 credits depending on compare + refresh

How It Works

  1. 1Audit your URL with target keyword (5cr) — or diagnose for site-wide work
  2. 2Optional: compare against a competitor URL (3cr)
  3. 3improve(freshness) / aeo_refresh for a section-by-section update plan (8cr), or improve(aeo) for rewrites

Agent Prompt Example

“Audit https://example.com/my-article for 'content marketing strategy', compare a competitor, and produce a freshness improve plan.”

8

Build a Content Strategy

Cluster keywords into pillars, then research and outline priority topics — or use aeo_create in strategy mode for the one-call composite.

Tool Chain

aeo_clusteraeo_create(strategy)aeo_researchaeo_outline

Cost

5 + (10–15 per pillar), or ~44 for create strategy mode

How It Works

  1. 1Cluster target keywords (5cr) — or aeo_create strategy / aeo_workflow_strategy (~44)
  2. 2Pick pillar topics with the human
  3. 3Per pillar: research → outline, then Agent Writes or hosted draft

Agent Prompt Example

“Cluster these 50 keywords for my SaaS blog, then research and outline the top 3 pillar topics.”

9

Citation & Visibility Monitoring

After publish: register URL track (Brave visibility proxy) and/or track questions on the citation board (measured ChatGPT / Perplexity / Gemini — page-tier green only). Never conflate the proxy with measured citations.

Tool Chain

aeo_trackaeo_scoreaeo_visibility

Cost

Track free; visibility/SOV 10cr when you run the proxy tool

How It Works

  1. 1Score content to establish a baseline (free)
  2. 2Register URL with aeo_track — Brave-search visibility proxy (not a measured engine citation)
  3. 3For retainer Monitoring: maintain branded vs goal questions on the citation board — green only on page-tier citations
  4. 4Read measured rates from GET /api/v1/visibility/scoreboard; use aeo_citation_sources / crawler-logs as needed
  5. 5Watch /monitoring for alerts (digest) — lost_citation on the proxy is not the same as losing a measured page citation

Agent Prompt Example

“I published https://example.com/my-article. Score it, register URL tracking, and tell me how to read the citation board vs Brave SOV honestly.”

10

Composite Workflows (Legacy Aliases)

Single-call composites still ship: audit_improve, strategy, health, quality_gate. Prefer aeo_create / aeo_improve / aeo_diagnose from onboard — they replace the common multi-step patterns with the same economics.

Tool Chain

aeo_workflow_audit_improveaeo_workflow_strategyaeo_workflow_healthaeo_workflow_quality_gate

Cost

0–44 credits (quality_gate free; strategy ~44; audit_improve ≤13; health 5/URL)

How It Works

  1. 1Prefer: diagnose for site audits, create for drafts/strategy, improve for rewrites, factcheck for stats
  2. 2If you already use composites: audit_improve for URL+refresh-below-threshold; strategy for cluster+pillars; health for batch audits; quality_gate for score pass/fail
  3. 3Check aeo_account / onboard for current credit costs before a large strategy run

Agent Prompt Example

“Prefer aeo_diagnose for my site audit. If you must use composites, run aeo_workflow_audit_improve on https://example.com/blog-post with keyword 'AI content strategy'.”

Tips

  • Prefer create / improve / diagnose / factcheck from GET /api/v1/onboard — granular tools stay as aliases.
  • Pass research between tools and into score — never skip research before write or submit-for-review.
  • Score after every edit — it's free. Use improve(to-score) instead of blind rewrites.
  • Measured citations ≠ Brave SOV proxy. Board green = page-tier only; never invent citation guarantees.
  • Check aeo_account before large strategy or study runs.