Advanced Mode — SEO Data Sheet, Outline & Generate

Advanced Mode gives you spreadsheet-level control over a bulk run. Each row is one article — with its own keyword, slug, title, title tag, secondary keywords, meta description, and context — and the flow moves through up to three steps:

The Outline step is optional — it only appears when you've enabled it in Settings → Prompts → Outline Step. With it off, the flow is just SEO Data → Generate.

Step 1 — SEO Data Sheet

Open the Generator, pick Advanced in the mode selector, and you land on the sheet. The toolbar shows live counts, search, a status filter, a Focus / Grid view switch, and Generate All SEO.

Focus view (above) pairs a scrollable row list with a full editor for the selected row — including a live meta-description meter that turns green in the ideal 150–160 range. Grid view shows every field in a dense spreadsheet table. The status filter (All / Ready / Draft / Done) and search box narrow the list.

Adding rows

Per-row Context (an optional field in the editor) steers research and writing for that row only — e.g. "Give a verified PSI range; cite the source; don't invent a precise figure." It's passed to the model as explicit instruction.

Generating SEO metadata

Generate All SEO (or the bulk-select bar's Generate SEO, or a row's Regenerate) fills in title, title tag, slug, secondary keywords, and meta description. On click you choose how existing values are handled:

• Fill empty fields only — keeps your hand-written titles/keywords and only populates blank fields (e.g. a missing meta). This is the safe default.
• Overwrite all fields — regenerates everything.
• Run as batch (~50% cheaper) — a Gemini-only toggle that submits the set as an async Batch job; rows fill in over a few minutes.

Speed: with Multi-Provider Parallel enabled, the run spreads across your provider lanes (e.g. Gemini ×3 + Claude ×1) at once. Otherwise it runs on your selected provider, in parallel across any multiple keys you've added.

Errors are surfaced — if a model rejects the request or returns non-JSON, the affected rows are tagged Error with a dismissible banner explaining why, and Generate All SEO re-runs error rows so you can retry after switching model/provider.

Row statuses

Step 2 — Outline (optional)

When the Outline step is enabled, Continue to Outline takes you to a stage that builds a per-row outline strictly from that row's SEO data + your chosen prompt + context — so the draft can't wander off-brief. Pick the outline strategy for the batch, then Generate All Outlines (or click any row to tweak its outline — editing is optional, so a 200-row run isn't gated on review).

• SEO outline — classic search-optimized H2/H3 structure.
• LLM / AEO outline — question/query-style headings tuned for retrieval and citation by AI answer engines.

Both prompts are editable (with Restore default) in Settings → Prompts → Outline Step, and outline generation supports the same Batch and Multi-Provider options. See the full Outline & AEO guide for the prompt mechanics.

Step 3 — Generate Content

The generate console replaces the old results list with a queue + live detail view. The top bar carries Start / Pause / Stop, progress, and (when done) Export CSV / ZIP and Publish all. Each queue item shows its phase while running and, when finished, its SEO score and lifecycle label:

Click any finished item to open the result detail — an SEO score ring, stats (words, headings, read time, citations), a content preview, and actions: Edit, copy markdown, and either Publish to WordPress or, once published, a green View live ↗ link straight to the post.

Publish lifecycle

Publish state and the live URL are saved to your project history, so the labels and links survive reloads. Live vs Published is driven by your project's WordPress default status (set it to Publish for true Live).

Scaling a run

Two complementary levers, mutually exclusive per run:

Multi-Provider applies to SEO, outlines, and drafting; Batch is offered for SEO and outlines.

Keyboard shortcuts

Shortcuts only fire when no input is focused, so they won't interfere with typing.

The same Generation Settings (model, word count, prompt mode, AI enhancements, content structure) described in Simple Mode apply to every row when you generate content — they're shared across modes.