The topic queue is your AEO content backlog, intent-classified, cluster-aware and de-duplicated before any article is drafted.
Topic queue
The topic queue is the ordered list of articles CiteFlow will generate next. Whether you generate manually or on a schedule, the next item in the queue is what gets written.
Where do topics come from?
Topics can enter the queue from four sources:
- KB recommended topics, auto-imported when you approve a knowledge base draft. These are the topics the synthesiser identified as gaps between your audit and your positioning.
- KB question patterns, the actual questions the synthesiser found prospects asking. Suggested alongside recommended topics.
- User-added, typed in directly, one at a time or via the multi-line bulk add textarea.
- AI-suggested, a separate "Suggest more" flow that asks the generator for fresh ideas given your KB and existing queue.
The source column on the queue tells you which path each topic came in by.
How do you manage the topic queue?
Open /dashboard/content and switch to the Topics tab. From there:
- Reorder, drag rows up and down. Order is persisted via the
display_ordercolumn and respected by the generator. - Skip, set a topic to
skippedto remove it from the next-up candidates without deleting it. Useful when a topic is timely later but not now. - Restore, bring a skipped topic back into the queue.
- Delete, remove permanently.
- Filter, narrow the view by source or status.
How do you bulk-add topics?
The bulk add textarea takes one topic per line and creates them in the order pasted. Useful when migrating from an existing editorial calendar or when an SEO consultant hands you a list.
- Empty lines are ignored.
- Each line is trimmed and de-duplicated against the existing queue.
How does a topic become an article?
When generation runs, either because you clicked Generate now or because the scheduler fired, CiteFlow:
- Picks the first non-skipped topic in queue order.
- Marks it as
in_progress. - Runs the generator using your approved KB as context.
- On success, attaches the new article to the topic and marks the
topic
completed. - On failure, marks the topic
failedwith the error and leaves the queue untouched so the next run picks it up again.
What's the difference between manual and scheduled generation?
- Manual, click Generate now on the topics page. Useful for one-off campaign articles or when you want to test a topic before committing it to the schedule.
- Scheduled, see auto-scheduling.
Both consume from the same queue, so a manually-generated article counts towards your tier limit just like a scheduled one.
What are the tier limits?
| Tier | Articles per month |
|---|---|
| Trial | 2 articles total over the trial period |
| Standard | 10 per month |
| Enhanced | 30 per month |
| Marketplace | Unlimited (soft fair-use cap) |
Limits reset at the start of each billing period. The topics page shows how many of your monthly allowance you've consumed.
How do you choose good topics?
Three patterns we see working:
- Question-led, phrase the topic as the question your prospect Googles ("how do I X", "what's the difference between X and Y"). Maps cleanly onto AEO-friendly structure.
- Decision-stage, write for the moment someone is deciding between options. These earn citations because they include comparisons, criteria, and trade-offs.
- Anchor terminology, use the named entities from your KB consistently. Repetition across your corpus is what makes LLMs connect the dots.
Avoid:
- Generic listicles ("Top 10 X tools") with no point of view.
- Topics that don't match your products or audience, the article will read as off-brand.
- Duplicates of pages you already have. Improve the existing page via the audit first.
How do you troubleshoot topic issues?
"My topics keep coming out as AEO-style articles even when I want a narrative piece."
CiteFlow's generator defaults to AEO structure (answer-first, scannable, with an FAQ section) because that's what earns AI citations. If you specifically want a long-form narrative, phrase the topic that way ("An essay on…", "Field notes from…") and the generator will adapt.
"A topic stayed in_progress for a long time."
Generation has a hard timeout. If a topic is stuck, refresh the page;
the worker recovers stuck jobs to failed after the timeout and the
next scheduled run will pick it up cleanly. You can also delete and
re-add the topic.
Related
- Knowledge base, the source of recommended topics and question patterns
- Article generation, what happens once a topic is picked
- Auto-scheduling, turn the queue into a publication cadence
References
- Semrush: keyword research guide, Semrush
- Google Search Central: structured data guidelines, Google
- Schema.org Article specification, Schema.org
Related
- Knowledge baseAuto-synthesised from your audit. Review, edit and approve before content generation reads it.
- Article generationClaude Sonnet generation, AEO structure, hero images, statuses, editing, approval and publishing.
- Auto-schedulingCadence, timezone, hour-only granularity, notification email, auto-publish and trial-expiry pause.