How Does the AI Welcome Post Draft Work During Onboarding?
During onboarding VeloCMS uses Gemini Flash to generate a first draft based on your blog's topic and intent — so you start with real content, not a blank page.
When you finish onboarding VeloCMS generates a draft welcome post using Gemini Flash. It's not a generic placeholder — the AI reads the blog name, topic, and audience description you provided during signup and writes something specific to your niche.
What information does the AI use to write the draft?
Three fields from your onboarding form feed the generation: your blog's name, the topic or niche you described, and the intended audience. The prompt is engineered to produce an answer-first opening paragraph (critical for AI search), two to three supporting sections with h2 headings, and a conversational closing. The generation runs in about eight seconds over a server-sent events stream — you can watch it appear word by word in the editor.
What if the draft isn't good?
You own the editor. Delete what you don't like, rewrite what's off, keep what works. The draft exists purely to kick-start momentum — the blank page is the enemy of getting started. You can also click Regenerate in the post's AI panel to get a completely fresh draft if the first one missed the mark. Regeneration uses the same onboarding context plus any edits you've made to the title, so iterating is quick.
The welcome post draft is created in your account automatically after onboarding success. It lands in your drafts, not published — you choose when (and whether) it goes live.
What happens if the AI generation fails?
If the Gemini API is unavailable or times out during onboarding, VeloCMS falls back to a short generic placeholder post rather than crashing the flow. You'll see a small notice in the editor letting you know the full AI generation didn't complete and offering a one-click retry. The fallback post is intentionally short so it's obvious you need to replace it — we don't want anyone accidentally publishing AI boilerplate.