Outerank Scribe
Your autonomous content agent — researches keywords, builds SERP briefs, writes articles. You approve, it publishes.
What it is
Outerank Scribe is a $19/mo add-on that runs the whole article workflow for you. In smart mode (the default), it discovers keyword candidates in your niche, validates them with real Google data (a 7B+ keyword index), picks a golden, not-yet-written target, builds a SERP brief from the top 10 ranking pages (intent, subtopics, key entities, top URLs), then writes a 1,500–2,500-word article on it — with internal-link suggestions, FAQ block, and JSON-LD schema. The draft lands in your Approvals queue with a 'How the agent researched this' transparency log so you see exactly what it looked at and why it picked that topic. You approve → it saves to your Content library and can be one-click published to WordPress / Shopify / Webflow / Ghost / HubSpot.
Why it matters
Doing real SEO content is 80% research and 20% writing. Scribe inverts the work: it does the research grunt (keyword discovery, SERP analysis, cluster shaping) and only asks you for the human-in-the-loop call: 'approve or reject.' One Scribe article costs ~$1 in real API spend; a freelance equivalent is $200+. And because it's transparent and human-gated, you keep editorial control without losing the throughput.
How to use it (step-by-step)
Follow these in order. Each step is 30 seconds to 2 minutes.
Step 1
Enable the add-on
Scribe is a $19/mo add-on. Once on, it appears in your sidebar as 'Outerank Scribe' with the Approvals queue.
Step 2
Pick a project + (optionally) add seed topics
Each project has its own queue. Add up to 24 topics in the keyword editor — each with a primary keyword + optional sub-keywords. Leave it empty and Scribe uses your project's niche instead.
Step 3
Choose smart mode (default) or manual
Smart mode (75 cr): Scribe discovers + validates + researches keywords on its own. Manual mode (50 cr): writes about the topics in your list, in rotation.
Step 4
Run on demand or schedule it
Click 'Run now' for one article right away, or check 'Run automatically on a schedule' and pick a cadence (twice-weekly / weekly / bi-weekly) to drip articles into your queue without lifting a finger.
Step 5
Approve or reject
Each pending draft has three read modes: 'Show its work' opens just the research log (fast confidence check), 'Preview full article' opens a proper modal for a real read-through with Approve/Reject in the footer, and 'Inline preview' shows a scrollable snippet in the queue. One click approves and drops the article straight into your Content library — a toast gives you a direct 'Open in library' link so you don't have to hunt for it. Reject and the keyword frees up for the next run.
Step 6
Publish when you're ready
Connect WordPress, Shopify, Webflow, Ghost or HubSpot once; publish any approved article to your site in one click. Decoupled from approval — no pressure to publish immediately.
Feed Scribe topics from Pulse — or pick manually.
Scribe writes on a schedule. Topics come from your latest Pulse brief automatically (the strategic intelligence agent feeding the writer), OR you pick topics manually from a list. Either way: zero copy-paste.
- Pulse-fed mode (auto)
- Manual-pick mode (you choose topics)
- Mixed mode (Pulse fills gaps in your manual list)
- Topic queue visible — no black-box scheduling
Scribe queue · this week
Same engine as Content — fully on its own.
Scribe uses the same SERP brief → article → coverage score pipeline as the Content module. The difference: Scribe runs unattended on your schedule, drops drafts into the approval queue, and waits for you to approve.
- SERP brief → article → coverage score pipeline
- 75 credits per article (vs 50 for manual Content)
- Schema + FAQ + meta included
- Tone-matched to your project's voice
Scribe draft · arrived
Draft · Title: "The Practical Guide to Leather Conditioning at Home"
Coverage · Score 84/100 vs live SERP — above median ranker.
Schema · Article + FAQPage JSON-LD attached.
One queue, one click.
Scribe queue lives at /modules/scribe. Each card shows the draft, its coverage score, the topic source. Approve → into the library. Reject → Scribe learns and skips similar topics next cycle.
- Single queue, every Scribe draft
- Approve / Reject / Edit on each card
- Reject teaches Scribe (it skips similar topics)
- Approved → searchable in your Content library
Approval queue
Publishing decoupled from approval.
Approving doesn't publish — it saves to your library. Publishing is a separate decision you control. Tracker auto-adds the keyword on publish. Brain measures rank movement 7, 14, 30 days post-publish.
- Approve ≠ publish (your call when to ship)
- Tracker auto-adds keyword on publish
- Brain measures lift 7d / 14d / 30d post-publish
- Failed-to-rank surfaces in the next Brain recommendation
Published · 30d performance
Why Outerank
What auto-blog tools forget about you owning.
Scribe writes; you own the voice, the editorial call, and the publishing decision.
Pulse-fed topic pipeline
Scribe doesn't need you to feed it topics — Pulse hands them over. The strategic-intelligence loop closes itself.
Same engine as Content
SERP brief → article → coverage score. Scribe is Content on autopilot, not a separate quality tier.
Reject teaches it
When you reject a draft, Scribe learns the rejection pattern. Next cycle, it skips similar topics.
Post-publish loop closes
Brain measures lift at 7d / 14d / 30d. Articles that didn't rank surface as recommendations to revise.
Tone-matched, not generic
If your project has a voice set (or learned), Scribe matches it. Not the standard AI-slop voice.
Approval gate, not raw auto
Scribe never publishes without your sign-off. The agent does the work; you keep the editorial call.
What costs credits
No surprises. Browsing and viewing saved results is always free.
Pro tips
- →Use the keyword editor's sub-keywords to nudge Scribe toward specific angles — they get woven into H2/H3s alongside whatever the SERP research surfaces.
- →The 'Research interrupted' step in the log is your friend — when an API hiccups, you'll see exactly which step failed and why, instead of getting a silent generic article.
- →Leave smart mode ON unless you have very strong opinions about exact-match keywords — the SERP brief makes the article rank for the whole cluster, not just one term.
- →Schedule weekly cadence and check the KPI strip at the top of the page — 'Next scheduled run in ~Xd' tells you the schedule is alive without opening the queue.
- →Word-count adherence got a bump 2026-07-12: two retries with widening tolerance (±150 → ±200 → ±250) so stubborn keywords stop shipping ±300-word articles.
Questions, answered
Site Audit FAQs
Will it always pick the same keyword?
What if every keyword is already covered?
What if generation fails after I get charged?
What if smart-mode research fails?
Are graphs and tables included?
Can I edit the title or body before approving?
How we compare
Honest comparison — where we win, where the big tools win.
vs. Surfer / Jasper
Both write — but they don't research, validate, or rotate topics for you. You hand them a keyword; Scribe finds the right keyword on its own and shows you the research.
vs. Hiring a freelance writer
Faster, cheaper, more transparent, and they never miss a deadline. The trade-off is you read the draft instead of handing notes.
vs. ChatGPT / the AI directly
You'd have to do the SERP research, build the brief, prompt-engineer the article, then format it. Scribe is the whole pipeline plus approval + CMS publishing.
Ready to try Outerank Scribe?
Open it on one of your projects.
Come back here whenever you need a refresher.