Frase Review 2026: My Hands-On Test From Onboarding to AI Visibility

I tested Frase in one hands-on session on 29 July 2026, on the 7-day trial that needs no credit card. I connected rankup.so and let Frase analyze the site before asking it for anything.

Frase categorized rankup.so as B2B SaaS for agentic SEO and GEO, then used that setup to research a topic and produce a content brief.

The writer produced a complete first draft, but the result still needed substantial work before publication.

This review follows that session stage by stage, then separates what I verified from the features Frase documents but I did not test.

My bottom line after one session with Frase

Frase moved quickly through site analysis, research, brief creation, and first-draft generation. The tradeoff was how much judgment the workflow left unresolved.

What worked in my test:

• Workflow continuity - Research, the brief, the draft, and optimization feedback stayed in one workspace.

• Useful constraints - The brief set a target length, reading level, tone, and audience before drafting.

Where the work remained:

• Publishability - The generated article lacked the product context, evidence, positioning, and voice needed for publication.

• Score interpretation - Frase's optimization checks surfaced useful gaps, but the scores could not establish accuracy, authority, or real AI visibility.

How this review was put together

The evidence comes from one hands-on session on 29 July 2026. I have not rechecked the product since, so documented capabilities are labeled separately from firsthand observations.

I tested on the Scale trial plan and entered rankup.so when onboarding asked for a domain to analyze.

What I tested firsthand:

• Trial signup and the site analysis of rankup.so

• AI Visibility prompt tracking across ChatGPT and Google AI

• The AI assistant chat, including a strategy question and a task run

• The Opportunities tab and the keyword clusters view

• Brief generation for the topic the assistant recommended

• AI drafting inside the editor

• The Content Diagnostic and its E-E-A-T scoring

• AI image generation

What I did not test:

• Market-wide keyword discovery and portfolio-level prioritization at scale

• Frase Answers and first-party question capture

• CMS publishing, Content Guard, and automatic publishing

• Google Search Console monitoring and content-decay tracking over time

• Long-term ranking outcomes

• Support, collaboration, and integrations

• Whether a brand profile improves output over repeated use

The site connected at onboarding was rankup.so, and the topic used for the full workflow was "8 best AI-powered SEO platforms for SaaS companies in 2026".

I assessed that draft rather than rewriting it, so what I describe is the state Frase delivered it in, not the article it could become after editing.

I checked pricing against Frase's own pricing page on the day of writing.

What is Frase in 2026?

Frase is a search-content platform covering topic research, briefs, drafting, optimization, CMS publishing, and post-publish monitoring.

Here is what every Frase plan now includes:

• Research, writing, and images - Topic research and SERP analysis feed a brief, an outline, an AI draft, and generated images.

• CMS publishing - Finished pages can go to WordPress, Webflow, Sanity, Wix, or the hosted FraseCMS.

• SEO and GEO scores - Content Score benchmarks search coverage, while the GEO score rates characteristics Frase associates with generative-engine readability.

• AI Visibility monitoring - This separate feature tracks whether supported AI engines cite your site. A high GEO score does not prove an AI citation.

• Content Guard - This watches published pages for Google rank or traffic decay and drafts a fix. Approval is the default, although users can opt into automatic publication for selected low-risk changes.

What happens at each stage of Frase's workflow?

The workflow runs in five stages: research and topic discovery, briefs and drafting, SEO and GEO optimization, publishing and post-publish monitoring, and AI visibility monitoring.

0. Account and brand setup

Setup matters here because Frase says it builds a brand profile from your existing content and carries that context into drafts. So what it picks up at this stage shapes every brief and draft that follows.

I ran this on a fresh trial account. Onboarding asked for a domain, a workspace name, and a starting point, where I picked AI visibility over a site audit.

That domain was the only brand input Frase got. From it, Frase read rankup.so and matched competitors on criteria like SEO workflow automation and topical authority mapping.

So the generic draft further down is what that setup produced. One session says nothing about whether Frase learns a brand over repeated use.

1. Research and topic discovery

Frase documents this stage as using the top 20 results to map competing coverage and topic gaps, with common questions and query-level keyword data alongside them.

What I saw firsthand was the setup around it. Frase let me review and adjust the topic, audience, and angle before research started.

When the assistant ran research on its own, it never surfaced those settings. I also opened the Opportunities tab and the clusters view under it:

• No keyword clusters were assigned in the report.

• The reset map control did nothing when I used it.

• Building the clusters looked like my own job.

So the workflow I ran started from a query I had already chosen. Nothing in it grouped the wider market into clusters for me.

Frase also documents Frase Answers and content-gap discovery as topic inputs. I did not test those features, so this limitation applies only to the query-led workflow I used.

2. Briefs, outlines, and drafting

Here is what that looked like in my session, in order:

1. I handed the AI assistant the writing task, and it ran research and generated a content brief.

   

2. It stopped at the brief. The article was not written until I explicitly told it to execute.

1. The writer showed "Preparing to write content...", then returned a finished draft in the editor.

   

Both the brief and the draft came back in Estonian, even though rankup.so, the blog I was writing for, and every message I sent the tool were in English. That was my biggest issue with the whole run.

The cause was a default language setting on the account, not a translation failure. Nothing in the workflow flagged it before a full 2,201-word draft came back in the wrong language.

The brief set these parameters:

• Word count - a 2,500-word target

• Reading level - intermediate

• Tone - authoritative

• Audience - SaaS founders and SEO specialists

The finished draft came back at 2,201 words, with 13 headers and 15 links, and a 73% combined optimization score shown beside them in the editor. It landed under the 2,500-word target the brief had set.

Hero image generation worked and returned a generic stock-style visual, a woman with a tablet and data charts.

When I asked for a body image instead, Frase proposed a Table Visualization based on the text rather than generating one.

The draft read like stitched-together SERP summaries. It carried no product knowledge, no angle of its own, no positioning and none of our voice.

3. SEO and GEO optimization

Frase’s SEO and GEO optimization is a draft-scoring workflow that compares content against top search results and flags keyword, topic, and structure gaps.

The finished draft received a 73% combined optimization score.

Those counts are what the score visibly keys off. It does not measure topical depth, competitive differentiation, or whether the article is useful to read.

Three things moved that number:

• Topic coverage - subtopics page one covered that my draft had not touched yet.

• Terms from ranking pages - words and phrases the top results used, listed so I could work them in.

• Structure and length - how my headings and word count compared with the competing articles.

The score worked as a pre-publication checklist, not a quality target. Adding length could raise it without making the draft more useful or accurate.

Clicking Improve EEAT in the sidebar opened a Content Diagnostic modal. It scored the same draft three ways and broke out Authority, Citable Data and Readability:

• E-E-A-T: 47 - Authority and Expertise were both flagged "Needs attention"

• GEO: 90

• SEO: 81

The Manual Tips accurately identified the missing citations, data, expert sources, and firsthand experience. Frase labeled the gaps but left them unresolved.

The GEO score rates generative-engine readability characteristics, while AI Visibility monitors actual citations. Frase does not explain how authority or visibility are calculated, so I treated the 90 as unvalidated.

No score in the editor or the diagnostic told me whether the draft was accurate or sounded like us. Both checks stayed manual.

4. Publishing, monitoring, and updates

Publishing, monitoring, and updates means getting content live, checking performance in Search Console, and revisiting pages that lose visibility or show topic and internal-linking gaps.

CMS publishing, auditing at volume, Content Guard, and Search Console decay tracking over time are documented features I did not run.

On paper this stage runs in three steps, none of which I ran myself:

1. Publish the finished page to WordPress, Webflow, Sanity, or Wix, or host it on FraseCMS.

2. Audit the pages you already have, scored for SEO, GEO, and AI visibility so the weakest ones surface first.

3. Watch live pages with Content Guard, which drafts a fix after detecting Google rank or traffic decay. Approval is the default, with optional automatic publication for selected low-risk fixes.

Content Guard watches 3 pages on Starter, 15 pages on Professional, and 50 pages on Scale. Audit volume runs from 50 pages a month to 1,000.

Search Console reporting has a delay. Treat a decay alert as a trend signal rather than a same-day verdict.

5. AI visibility monitoring

Frase's AI Visibility module reports whether the AI engines included in your plan mention your site in answers to the prompts you track. Every number it shows is calculated from that prompt list.

My workspace opened with two prompts Frase had written for itself during setup. Both were broad category questions rather than the SaaS-specific queries rankup.so would actually compete for:

Those two prompts were the entire basis for the first report. The Overview tab scores this set of metrics from whatever you track:

• Share of voice - your share of total brand mentions compared with competitors

• Sentiment - how the AI platforms describe your brand

• Appearance rate and authority rate - how often your site turns up in answers, and how prominently

• Trend velocity and momentum - whether those numbers are moving

• Appearance rate over time - the same appearance data plotted as a chart

Every card came back at 0% for rankup.so, and the appearance-rate chart was an empty axis:

The Opportunities panel on the same screen disagreed with the cards above it. It read "No gaps found. You are being cited everywhere competitors are" while both tracked prompts sat at Not visible:

To get past two prompts I used Frase's own suggestions, which it grouped by where a buyer sits in the journey:

• Awareness - top-of-funnel discovery queries

• Consideration - solution-seeking queries

• Decision - competitive comparisons

That second batch was sharper than the two I started with. "AI visibility tracking 2026" and a question about optimizing SaaS blog content were both close to what my buyers would type.

I could only add three of them. The trial capped the workspace at five tracked prompts, and going past that meant upgrading:

Each prompt carries four controls: edit, pause tracking, archive, and run now.

Running all five returned "visibility triggered results will show shortly". Every one resolved to Not visible, with the change listed as stable.

A stable trend is doing strange work when the number underneath it is zero, and nothing in the interface explained what that label measures.

The cap turned up again on the Alerts page as a "5 / 5 prompts used" warning. That meter is separate from the plan's AI generations allowance, so how much of the market you can watch is metered on its own:

Two things limited what I could do from there:

• No alerts of my own - I could not find a way to create one. The assistant confirmed alerts fire automatically on a ranking drop, a competitor gain, or a lost citation, and cannot be set up by hand. Its reply came back in Estonian, the same account default that produced the draft.

  

• No execution path - Asked what to do with the visibility data, the assistant reported zero citations across the tracked engines and recommended GEO work plus articles aimed at specific queries. Nothing in the module carries that out.

  

Reporting is where this stage stops. Moving a prompt off Not visible meant going back to the editor, where the draft quality problem from the drafting stage was still waiting.

What should you know about Frase pricing?

Frase pricing runs from $39 to $239 per month on annual billing across three self-serve plans, Starter, Professional, and Scale, with custom pricing for Enterprise.

Older reviews may still mention Solo, Basic, and Team. Frase now lists Starter, Professional, Scale, and Enterprise plans.

Your actual monthly cost depends on seats and optional overages:

• Extra seats - Professional and Scale charge $29 per month for each additional seat.

• Optional overages - Professional charges $5 per article, $0.50 per audit page, and $0.25 per AI prompt. Scale charges $4, $0.40, and $0.20 respectively. Pay-as-you-go is off by default.

• Annual billing - Paying yearly reduces the month-to-month rate by 20 percent.

• Trial - The standard seven-day trial uses the Professional feature set and caps usage at 5 articles, 50 audit pages, 25 AI prompts, and up to 3 seats. Auto-publish and other agent autonomy are disabled.

Editorial time still affects the value of an article allowance. That calculation may look different if you also use Frase for audits, monitoring, publishing, or Content Guard.

Frase calls the included allowance “AI generations” in its comparison table and “AI prompts” in the overage FAQ. Plans otherwise differ by volume, team size, monitoring coverage, and create-side autonomy.

Who is Frase best for, and who should skip it?

Frase is easiest to justify when its research, drafting, optimization, and post-publish tools replace several separate tasks.

Frase is a practical fit when you need:

• A document-centered workspace for SERP research, repeatable briefs, and guided optimization

• A fast first structure and are comfortable reviewing the draft for accuracy, product detail, and voice

• Auditing, CMS publishing, AI Visibility, or Content Guard alongside new-article production

• Consistent briefs across a solo workflow or small content team

Frase is a weaker fit when you need:

• A brand-ready draft with little setup or editorial review

• Market-wide keyword work - I did not test clustering or portfolio-level prioritization, so plan on another tool for that layer.

• A GEO score you can trust on its own - The GEO number came back high while AI Visibility reported zero citations for rankup.so, so I could not treat it as a standalone signal.

• A firsthand verdict on publishing, monitoring, support, or collaboration

Low publishing volume does not automatically make Frase poor value. A team maintaining a large existing library may get more from auditing and monitoring than from the monthly article allowance.

How does RankUp compare as a Frase alternative?

RankUp is the alternative for SaaS teams that want strategy, writing, and updates carried through with persistent business context. Its SEO and GEO agents start with your brand voice, product knowledge, market research, and approved interview answers already available.

That shared context supports three distinct roles:

1. Strategy. Magnus discovers competitors and keywords, clusters them into topics, and sets the priority before any drafting starts:

1. Writing. Cedric takes that prioritized topic into live SERP research, then builds the outline and blueprint the draft is written against. He asks focused questions only where the knowledge base has a gap:

1. Updates. Lyra audits live pages, decides what each one needs, and routes the writing to Cedric for review. She works from the same knowledge base rather than a fresh prompt:

That is the practical distinction. Frase documents a broad content loop, while RankUp is built around a SaaS-specific team whose context compounds across planning, writing, and updates.

Frase is worth considering for fast research, repeatable briefs, and guided SEO content optimization. The main limitation in my session was not feature availability, but the editorial work needed to make the draft accurate, specific, and recognizably ours.

If those editing and context handoffs are the work you want removed, test the alternative on one real article. You can start trial for $25 with an initial brand voice, knowledge base, and topical map built from your site, plus $50 in agent credits.

More questions about Frase

Is Frase worth it?

Frase is worth considering if fast SERP research, repeatable briefs, guided optimization, and the wider publishing and monitoring loop replace meaningful work for you. It is harder to justify if you expect a brand-ready draft with little review.

Does Frase have a free trial or free plan?

Frase offers a 7-day free trial with no credit card required, and there is no permanent free plan.

The standard trial lasts seven days and requires no credit card. It uses the Professional feature set with caps of 5 articles, 50 audit pages, 25 AI prompts, and up to 3 seats.

Auto-publish and other agent autonomy are disabled during the trial. Paid plans start at $39 per month on annual billing.

How does Frase billing work?

Frase billing is a tiered subscription charged monthly or yearly, with per-plan caps on articles, audit pages, seats, and AI actions.

Self-serve plans range from $39 to $239 per month on annual billing, or $49 to $299 month to month. Enterprise pricing is custom.

Annual billing is 20 percent lower than the month-to-month rate. Professional and Scale include optional pay-as-you-go usage and charge $29 per month for extra seats; upgrades are prorated, while downgrades begin next cycle.

What happens if you hit plan limits?

Hitting Frase plan limits means your article, audit page, seat, or AI action cap applies until you upgrade or turn on pay-as-you-go.

Starter stops at its monthly limit until the cycle resets or you upgrade. Professional and Scale can continue only if you enable optional pay-as-you-go usage.

Overages cover articles, audit pages, and AI prompts, while extra seats cost $29 per month. The pricing section above lists the plan-specific rates and included monthly allowances.