Keyword-led topic research
Candidate topics come back ranked by search intent and content gap, each with a primary keyword, secondaries, and a stated reason it is in demand right now.
Google-policy safe by design
Research, write, illustrate and publish 3,000 to 4,000 word posts that read like a human wrote them, tuned to your industry instead of a generic template.
Active industry profile: DoTechTalk (Tech News & Education Blog)
Pick one or more — each selected topic is queued as its own background post.
| Title | Keywords | Format | Why now / why it ranks here | |
|---|---|---|---|---|
| How to spot a fake crypto airdrop before you connect a wallet | fake airdrop warning signs wallet drainer, crypto scam checklist | Troubleshooting Guide | Airdrop-farming searches spike after every major token launch | |
| Web3 wallets vs custodial exchanges: who should hold your keys? | web3 wallet vs exchange self-custody, hot vs cold wallet | Comparison Post | Renewed self-custody interest after recent exchange outages | |
| What is a smart contract, in plain English | what is a smart contract smart contract example, how it works | Glossary Definition | Steady definitional volume, fills a gap in the beginner cluster | |
| The 11-point checklist before you mint your first NFT | nft minting checklist gas fees, minting cost | Checklist | Beginner minting queries rising with new low-fee chains |
Timezone
IST — Mumbai, Delhi, BangalorePublishing days
Word count
Writing tone
ConversationalFeatures
Inside the plugin
Every post runs through research, a word-budgeted outline, section-by-section writing, illustration and a compliance check before it ever reaches your editor.
Candidate topics come back ranked by search intent and content gap, each with a primary keyword, secondaries, and a stated reason it is in demand right now.
Every topic gets the shape that fits it — comparison, checklist, glossary, playbook, myth-buster — modelled on how Stripe, HubSpot and Ahrefs actually structure posts.
A word-budgeted outline first, then each section written and re-checked against its own target. No 1,500-word draft pretending to be a 2,500-word one.
Tone, vocabulary, audience and topic focus load from a profile per site, so a hemp dispensary and a B2B SaaS never end up sounding like the same writer.
Banned claims are matched in code, not just requested in the prompt. A hit holds the post for review even when auto-publish is switched on.
Em-dashes stripped, 30+ AI clichés blocked, sentence length forced to vary — the specific patterns detection tools score against.
Anchored table of contents, FAQ accordion with FAQPage schema, internal links placed inline, and title, meta and focus keyword written straight into Yoast.
A featured image plus inline visuals only where the outline marked a genuine need for one, generated to match the image style you picked.
Post counts and times per day in your own timezone, generated in the background so a slow run never times out your browser or your host.
How it works
One call to the plugin runs all five. You see the topic list and the finished post; everything between them is handled.
01 Research and rank topics
| Title | Primary keyword | Format |
|---|---|---|
| Web3 wallets vs custodial exchanges: who should hold your keys? | web3 wallet vs exchange | Comparison Post |
| How to spot a fake crypto airdrop before you connect a wallet | fake airdrop warning signs | Troubleshooting Guide |
| What is a smart contract, in plain English | what is a smart contract | Glossary Definition |
Each topic carries its own secondary keywords and a stated reason it is in demand, so you can reject one before it costs you a generation.
02 Budget the outline
Budgets are planned about 12% above your floor, because sections undershoot their own target. A table or a visual is only flagged where the content genuinely has one.
03 Write section by section
Section three came in short, so only section three is rewritten with more depth. The other 2,400 words are never regenerated, and never re-billed.
04 Illustrate and assemble
| Self-custody | Exchange | |
|---|---|---|
| Who holds the keys | You | The platform |
| Recovery if lost | Seed phrase only | Support ticket |
Anchored contents, stat bars, styled tables, an FAQ block with FAQPage schema and two CTAs, plus title, meta and focus keyword written into Yoast.
05 Publish, or hold it back
| Post | Status |
|---|---|
| Web3 wallets vs custodial exchanges | Published |
| What is a smart contract, in plain English | Published |
| Why this token could double by December Matched a banned claim: price predictions stated as fact | Held for review |
The guardrail is code, not a prompt instruction. It overrode auto-publish and sent that post to pending instead, which is the difference between a policy problem and a draft.
Reads human
Detection tools do not read for meaning. They score patterns: an em-dash habit, a predictable cliché vocabulary, and sentences that all run the same length. Each of those is handled in the prompt, then checked again in code after the draft comes back.
See where it happens
The active profile's banned claims and required disclaimers go into every section prompt, before a word is written.
When the draft comes back, the finished body is matched against that same list. A prompt can be ignored by the model. This check cannot.
A match forces the post to pending and records exactly which claim it hit, even with auto-publish switched on.
Compliance
A hemp dispensary cannot say FDA-approved. A crypto blog cannot promise a return. Each profile carries its own banned claims and required disclaimers, and they are enforced against the finished draft rather than politely requested in the prompt.
See the profiles3 more claims, plus 5 disclaimers appended automatically
Content formats
A checklist does not read like a comparison, and neither reads like a glossary entry. The research step picks the format that fits each topic, so a year of posts does not come out looking like the same article rewritten twenty times.
See how topics are picked
The full library
The last three are switched off for auto-generation on purpose. Each one rests on something the plugin does not have: a real survey, a real named interviewee, or a real internal experiment. Producing them anyway would mean inventing the evidence, which is worse than not publishing them.
Industry profiles
A profile is a single JSON file. It sets the tone, the vocabulary, who is being written for, which topics matter, and the claims that must never appear. Five ship as working examples, including a regulated hemp retailer and a crypto news site. Any other industry is a new file, not a code change.
What a profile controls
Pricing
No OpenAI or Gemini account to open, no API keys to paste into a client's site, and no separate token bill arriving at the end of the month. Writing and images run on our keys.
One site, a steady post every few days, written while you do something else.
$ 29 / month
30 credits a month
For agencies running several client blogs, each in its own voice.
$ 79 / month
100 credits a month
For networks and publishers pushing volume across many sites.
$ 199 / month
300 credits a month
One credit is one finished post: researched, written to 3,000-4,000 words, illustrated with two images, and checked against the rules of your industry. A credit is spent when the post is written, not when you publish it, so a post held for review still counts. A run that fails before producing a draft does not.
Built against real constraints
No testimonials here yet, because there are none worth quoting yet. What there is: four working profiles, each written against a real business with rules a generic AI writer would have walked straight through.
Cannot say FDA-approved, cannot call Delta-8 non-psychoactive, cannot promise a post-use drug test comes back clean. Nine banned claims and five disclaimers, enforced on the finished draft.
No price predictions stated as fact, no guaranteed returns, no implying the site gives licensed financial advice. The exact category the FTC and SEC pursue publishers over.
Insurance-agency vocabulary, and a standing rule against name-dropping the market-leading rival as the go-to example when a generic phrase does the same job.
A product that sells on transparency rather than surveillance, so the writing had to stay on the right side of that line without being reminded every time.
We’ve got answers
The questions people actually ask before buying: what it costs to run, what happens on publish, and what the plugin refuses to do.
No. Writing and images run on our keys, so there is no API account to open and no separate token bill. You buy credits, one per post, on the plan you pick.
Google's policy targets scaled content abuse: automated publishing at volume with no human review. That is why auto-publish ships switched off, every post lands as pending for you to read, and a post that trips a compliance guardrail is held back even when auto-publish is on.
Usually one to three minutes. It runs in the background through WP-Cron rather than inside the page request, so the button returns immediately and a slow generation can never time out your browser or your host.
3,000 words is the floor and 4,000 the ceiling, both adjustable. Actual length is decided per post by what the topic needs. The outline gives every section its own word budget, and a section that comes in short is rewritten on its own rather than the whole draft being regenerated.
Yes. Run just the research step and you get candidate topics with the keyword, the format picked for each, and why it is in demand right now. Tick the ones you want and leave the rest, or let it choose the top-ranked topic itself on a schedule.
Yes. If Yoast is active, the title, meta description, focus keyword and social fields are written straight into Yoast's own meta keys, so its UI and analysis pick them up with no extra step. If it is not installed, the plugin outputs those tags itself.
The claim is matched against your profile's banned list in code after the draft comes back, not just requested in the prompt. A hit forces the post to pending and records which claim it hit, so you see exactly why it stopped.
No. The research step picks a format per topic from a library of 24, and the outline plans different sections, tables and visuals for each. Titles are forced to vary in phrasing across a batch too, so you do not end up with twenty posts sharing one headline shape.
That is what the plans are built around. Each site gets its own industry profile, so a hemp retailer and a B2B SaaS never sound like the same writer, and every site draws on one shared pool of credits rather than a separate bill each.
Ready when you are
Two posts a week, in your industry's voice, held for your review before anything goes live.