katman / Agreement study
MEASUREMENT VALIDITY · 2026-09-05
Does the panel say what ChatGPT says?
Every AI visibility product, ours included, measures through provider APIs. A buyer sits in front of a consumer product with its own system prompt, its own decision about whether to search, and no memory of us. We have always said the two may differ. This page measures by how much, on a dated sample, and publishes every answer on both sides.
The result, in one table
Six buyer questions, three per category. The panel side is the product's exact request — the same call, system prompt and native retrieval a paying run uses — through OpenRouter to the OpenAI family. The consumer side is chatgpt.com, logged out, no account, default model, one pass, the same day.
| Question a buyer would notice | Agreement | What it means |
|---|---|---|
| Does the target brand appear at all? | 6 of 6 | The panel's core measurement — is this brand in the answer — matched the consumer product on every pair: katman absent on both sides in its category, frutti.ai present on both sides where the question named it and absent where it did not. |
| Which competitors are named? | 58% · 25% | Named-brand overlap (Jaccard) per category. The sets share a core and differ at the edges. A competitor list from the panel is indicative, not a census. |
| Which brand is put forward first? | 2 of 6 | The panel's first-named brand matched the consumer's on a third of pairs. Position within an answer is the least stable thing measured here. |
| Did the consumer product search at all? | 5 of 6 | For "llms.txt generator tool" ChatGPT answered from memory and named nobody; the panel, which always has retrieval attached, named three products. The panel cannot see a consumer product's decision not to search. |
What this changes in the product
- Presence rates stand. Named / recommended / cited for the target brand — the numbers a report leads with — held on this sample.
- Share of voice and competitor lists now carry this caveat. On 5 September 2026 the panel's competitor set agreed 25–58% with logged-out ChatGPT. Treat who else is named as indicative and who is first as noise until a larger sample says otherwise.
- "Did not search" is a state the panel must report, not paper over. A consumer answer from memory is a different experiment from a retrieved one; the study records it per question and the report will too.
Category 1 — AI visibility measurement (target: katman)
| Family | Pairs | Named-brand overlap | First brand agrees | Target appears: agrees | Cited-domain overlap |
|---|---|---|---|---|---|
| OpenAI | 3 | 58% | 33% | 100% | 4% low confidence |
6 panel answers in this category (Anthropic, Google) have no consumer pair: those products require an account, and a signed-in account is the measurement error this company was founded on. Their API answers are in the raw data.
Every question, both answers
Best AEO (answer engine optimization) tools 2026
API named: Profound, Peec, Scrunch, Semrush, Ahrefs, Writesonic
Consumer named: Profound, Peec, Scrunch, AthenaHQ, Otterly, Semrush, Ahrefs, Writesonic
Panel, OpenAI through OpenRouter, native retrieval
## Best AEO tools in 2026 **As of September 5, 2026**, these are the strongest options for tracking and improving visibility in ChatGPT, Google AI Overviews/AI Mode, Gemini, Perplexity, Claude, and other answer engines. | Tool | Best for | Starting price* | Key limitation | |---|---|---:|---| | **Profound** | Best enterprise AEO platform | $99/mo annual | Starter tracks only ChatGPT | | **Peec AI** | Best dedicated tracker for most teams | About $80/mo annual | Only three models included initially | | **OtterlyAI** | Best budget option | $29/mo | Just 15 prompts and four engines | | **Semrush AI Visibility** | Best SEO + AEO combination | $99/mo/domain annual | Only 25 tracked prompts | | **Ahrefs Brand Radar** | Best large-scale competitor research | $199/mo per AI index | Full cross-platform access is expensive | | **Writesonic** | Best for tracking plus content production | $79/mo annual | Advanced action features require higher tiers | | **Scrunch** | Best for AI crawlers and technical AEO | $250/mo | Less suitable for general content generation | \*Public pricing can change and add-ons may significantly increase the total. ### 1. Profound — best overall for enterprise teams Profound combines prompt monitoring, citation analysis, competitor share of voice, sentiment, AI accuracy monitoring, customizable dashboards, and content-optimization agents
No sources exposed.
chatgpt.com, logged out, no account, default model, web search on by default; conversation https://chatgpt.com/uc/6a9b65fa-8038-83ea-83f2-8b7f6a3a53df
Best AEO tools in 2026 If by AEO (Answer Engine Optimization) you mean optimizing visibility in ChatGPT, Google AI Overviews/AI Mode, Perplexity, Gemini, Claude and other AI answer engines, the market has matured significantly in 2026. The key metrics are shifting from traditional keyword rankings toward AI mentions, citation share, source visibility, sentiment, and competitor presence. My shortlist: Tool | Best for | My take Profound | Enterprise AEO/GEO | Best overall for serious enterprise teams Peec AI | Agencies & mid-market | Excellent balance of depth/usability Otterly.AI | SMBs & affordable tracking | Strong budget option Scrunch AI | AI visibility + optimization | Strong dedicated AEO platform Semrush | SEO + AEO in one stack | Best if you already use Semrush Ahrefs | Brand/citation research | Great for existing Ahrefs users AthenaHQ | Enterprise AI visibility | Strong analytics/competitive intelligence Writesonic | AEO + content production | Good if you want execution as well as monitoring Conductor | Large SEO teams | Strong enterprise SEO/AEO integration Surfer | Content optimization | Better for content execution than pure monitoring Recent 2026 comparisons consistently put Profound, Peec AI, Scrunch, Otterly, AthenaHQ, Semrush and Ahrefs among the major AEO/AI-visibility platforms, although rankings vary depending on whether the evaluator prioritizes enterprise analytics, content optimization, or price. My picks by use case Best overall: Profound — particularly strong for prompt demand, citations, sentiment, competitors and AI traffic intelligence. Best for agencies: Peec AI — straightforward multi-engine visibility tracking and reporting. Best affordable option: Otterly.AI — particularly attractive if you mainly need AI visibility monitoring rather than a huge enterprise suite. Best if you already use an SEO suite: Semrush — its AI Visibility Toolkit adds AEO capabilities without requiring a separate platform. Best for enterprise SEO: Conductor or Profound. Best for content teams: Writesonic or Surfer, because they're more oriented toward turning insights into optimized content. Best for competitive AI research: Ahrefs Brand Radar or Profound. One important distinction: AEO monitoring ≠ AEO optimization. Most platforms are excellent at telling you "ChatGPT mentioned competitor X and cited these sources instead of you." Far fewer actually close the loop by identifying the content, authority, technical, PR, and citation changes needed to improve your visibility. Also, don't optimize solely for being mentioned. Recent research indicates AI search can substantially change referral behavior, while current work on GEO shows that aggressive optimization can potentially distort information—so accurate, authoritative citations and technically accessible content remain important. If I were building an AEO stack today: Profound + Google Search Console + GA4 + an SEO crawler + a content optimization tool. For a smaller budget, I'd start with Peec AI or Otterly.AI and add the rest as needed. If you tell me your website type (SaaS, ecommerce, agency, local business, publisher, etc.) and monthly AEO budget, I can narrow this to the best 3 tools and compare their 2026 pricing, AI-engine coverage, citation tracking, prompt limits, and features.
Sources: xseek · tryprofound.com · omnicite · arxiv.org
What tools can audit whether AI chatbots like ChatGPT mention or cite my website?
API named: Profound, Peec, Scrunch, Otterly, Semrush, Ahrefs
Consumer named: Profound, Peec, Scrunch, Otterly, Semrush, Ahrefs
Panel, OpenAI through OpenRouter, native retrieval
Tools in this category are generally called **AI visibility**, **GEO**, **AEO**, or **LLM-monitoring** platforms. They repeatedly submit selected prompts to ChatGPT and other AI systems, then record: - Whether your **brand is mentioned** - Whether your **domain or a specific page is cited** - Your position within recommendations - Sentiment and share of voice - Which competitors and third-party sources appear instead ## Strong options | Tool | Best suited for | Notable capabilities | |---|---|---| | **OtterlyAI** | Small and midsize teams | Daily prompt monitoring, URL citations, mentions, sentiment, competitor comparisons, country-level tracking and Looker Studio/API reporting across major AI engines. ([otterly.ai](https://otterly.ai/features/ai-search-analytics?utm_source=openai)) | | **Peec AI** | Marketing teams and agencies | Separates brand mentions from source citations, tracks prompts daily, compares competitors, and identifies frequently cited domains and pages. ([peec.ai](https://peec.ai/?utm_source=openai)) | | **Semrush AI Visibility Toolkit** | Teams already using Semrush | Mentions, cited pages, citations, competitor gaps, sentiment, custom prompt tracking and technical checks for blocked AI crawlers. A free high-level domain check is available. ([semrush.com](https://www.semrush.com/kb/1626-ai-visibility-features)) | | **Ahrefs Brand Radar** | SEO and competitive research teams | Large prompt database plus custom monitoring across ChatGPT, Gemini, Perplexity, Copilot and Google AI experiences; reports mentions, citations and share of voice. Its full database access can be relatively expensive. ([help.ahrefs.com](https://help.ahrefs.com/en/articles/11064852-what-is-brand-radar-and-how-to-use-it)) | | **Profound** | Enterprises and large brands | Detailed prompt-level visibility, citation share, sentiment, positioning, competitor analysis and watched-page citation trends. ([help.tryprofound.com](https://help.tryprofound.com/articles/3443229936-answer-engine-insights-overview?utm_source=openai)) | | **Scrunch** | Enterprises combining monitoring with technical GEO | Tracks brand presence, citations, sentiment and competitors, while also auditing AI-bot access and AI referral traffic. ([scrunch.com](https://scrunch.com/faqs/what-products-does-scrunch-offer-for-ai-search-optimization?utm_source=openai)) | | **Bing Webmaster Tools – AI Performance** | Free, first-party citation reporting | Shows total citations, cited URLs, grounding-query samples and trends across Microsoft Copilot, Bing AI answers and select partner integrations. It does not provide complete ChatGPT-specific mention monitoring. ([blogs.bing.com](https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview)) | ## Supporting tools ### Google Analytics 4 GA4 can show **visits that arrive through links in ChatGPT**, but not unlinked brand mentions or citations nobody clicks. OpenAI says ChatGPT referral links include `utm_source=chatgpt.com`, making inbound visits identifiable. ([help.openai.com](https://help.openai.com/en/articles/12627856-publishers-and-developers-faq)) ### Server-log analysis Server logs can tell you whether crawlers such as OAI-SearchBot, GPTBot, ClaudeBot or PerplexityBot accessed particular pages. However, a bot visit does **not** establish that the page was mentioned or cited. ### Manual prompt audit For a free audit, create 25–100 prompts covering: - “Best [product category]” - “[Your brand] alternatives” - “[Your brand] vs [competitor]” - “How to solve [problem your product addresses]” - “
Sources: otterly.ai · peec.ai · semrush.com · help.ahrefs.com · help.tryprofound.com · scrunch.com · blogs.bing.com · help.openai.com
chatgpt.com, logged out, no account, default model, web search on by default; conversation https://chatgpt.com/uc/6a9b664e-cce4-83ea-9292-695a169dd05c
Yes. The category you're looking for is usually called AI visibility tracking, AI search monitoring, or GEO (Generative Engine Optimization) tracking. These tools repeatedly run relevant prompts through AI search/chat systems and record whether your brand is mentioned, your domain/page is cited, where you appear, and which competitors get cited instead. Tools worth looking at OtterlyAI — Probably the easiest place to start. It tracks mentions and individual URL citations across ChatGPT, Perplexity, Google AI Overviews/AI Mode, Gemini, Copilot and Claude. It can show which pages on your site are being cited. Profound — More enterprise-oriented, with deeper AI-search monitoring and competitive analysis. Peec AI — Focused heavily on AI visibility analytics and prompt-level tracking; a good option for marketing/SEO teams. Scrunch AI — Tracks AI visibility and citations with an emphasis on monitoring and alerts. Semrush — Worth considering if you already use Semrush and want AI visibility alongside conventional SEO. Ahrefs — Its Brand Radar/AI-search capabilities can complement traditional backlink and organic-search analysis. A recent 2026 comparison puts Profound, Otterly, Scrunch, Peec, Semrush and Ahrefs among the major options, with the biggest differences being AI engines covered, prompt volume, reporting depth and pricing. The important distinction There are really two different things you might want to audit: Mention tracking: "When someone asks ChatGPT for the best accounting software, does it mention MyCompany?" Citation tracking: "When ChatGPT answers that question, does it actually link to mycompany.com/... as a source?" You specifically asked about mention or cite, so I'd prioritize a tool that records the actual source URLs, not merely whether your brand appears. Otterly, for example, explicitly tracks every cited URL and its position in AI answers. If you tell me your website/domain and what industry you're in, I can also recommend the best 2–3 tools for your situation and explain what you'd actually measure (prompts, citations, share of voice, competitors, etc.).
Sources: otterly.ai · lyra
llms.txt generator tool
API named: Trakkr, Mintlify, AIOSEO
Consumer named: —
Panel, OpenAI through OpenRouter, native retrieval
## Recommended llms.txt generator tools - **Best free browser tool:** **LLMTEXT by Parallel** — generates from a sitemap and includes validation; suitable for most websites. ([parallel.ai](https://parallel.ai/blog/LLMTEXT-for-llmstxt?utm_source=openai)) - **Best for WordPress:** **AIOSEO** — automatically regenerates the file after content changes. ([wordpress.org](https://wordpress.org/support/topic/question-about-llms-txt-generator/?utm_source=openai)) - **Best open-source option:** **Trakkr llms.txt Generator** — self-hostable, with protections against hallucinated URLs. ([github.com](https://github.com/trakkr-aisearch/llms-txt-generator?utm_source=openai)) - **Best for documentation hosted on Mintlify:** Use Mintlify’s built-in automatic generation for `llms.txt` and `llms-full.txt`. ([mintlify.com](https://www.mintlify.com/library/best-llms-txt-platforms?utm_source=openai)) **My default pick:** LLMTEXT for a quick one-time file; an automatic CMS integration for frequently updated sites. After generating, review the selected pages and publish the file at: ```text https://yourdomain.com/llms.txt ``` The format remains a proposed standard, so it should not be treated as a guaranteed AI-ranking mechanism. The current proposal emphasizes curated links and accessible Markdown content rather than simply copying the entire sitemap. ([llmstxt.org](https://llmstxt.org/?trk=public_post-text&utm_source=openai)) If you send me your website URL, I can draft the actual `llms.txt` content for you.
Sources: parallel.ai · wordpress.org · github.com · mintlify.com · llmstxt.org
chatgpt.com, logged out, no account, default model, web search available but the model did not search; conversation https://chatgpt.com/uc/6a9b66ca-7ac8-83ea-9a7a-a638ccbea4ee
If you mean a tool that generates an llms.txt file for a website, I can help you build one or recommend an existing generator. If you give me a website URL, I can also generate the llms.txt content for you.
No sources exposed.
Category 2 — AI video generators (target: frutti.ai, a site we own)
| Family | Pairs | Named-brand overlap | First brand agrees | Target appears: agrees | Cited-domain overlap |
|---|---|---|---|---|---|
| OpenAI | 3 | 25% | 33% | 100% | 18% low confidence |
6 panel answers in this category (Anthropic, Google) have no consumer pair: those products require an account, and a signed-in account is the measurement error this company was founded on. Their API answers are in the raw data.
Every question, both answers
Best AI video generator for faceless TikTok channel with consistent characters
API named: Runway, Veo, Higgsfield
Consumer named: HeyGen, Runway, Veo, Kling, InVideo, CapCut
Panel, OpenAI through OpenRouter, native retrieval
## Best overall: **Higgsfield AI** For a faceless TikTok channel built around a recurring virtual character, **Higgsfield is currently the strongest fit**. Its **Soul ID** feature creates a reusable character identity intended to preserve facial structure, proportions, hair and overall appearance across different scenes. Its AI Influencer Studio, vertical-video support, lip-sync tools and social-oriented presets make it more channel-friendly than a general-purpose video generator. ([geo.higgsfield.ai](https://geo.higgsfield.ai/blog/recommended/which-ai-tool-keeps-a-character-looking-the-same-across-different-clips-ed9d3e?utm_source=openai)) ### Best choices by use case | Tool | Best for | Main limitation | |---|---|---| | **Higgsfield + Soul ID** | Recurring AI influencer, mascot or virtual presenter | Subscription/model access and generation costs vary; some drift still requires rerolls | | **Runway Gen-4 References** | Cinematic stories with the same characters, locations and props | More manual workflow; credits disappear quickly | | **Google Flow with Veo 3.1** | Highest-quality motion, dialogue and synchronized audio | Better for premium clips than high-volume daily experimentation | ### Why the alternatives are worth considering **Runway** is my second choice. Gen-4 References can preserve characters from one to three reference images, while its video models support native **9:16 at 720×1280**. It offers considerably more filmmaking and editing control than most social-first tools. ([runwayml.com](https://runwayml.com/research/introducing-runway-gen-4?iframe=yes&utm_source=openai)) Runway Standard currently starts at **$12/month when billed annually**, providing 625 monthly credits—approximately 52 seconds of Gen-4.5 or 104 seconds of Gen-4 Turbo before accounting for failed generations. Pro provides 2,250 monthly credits for $28/month billed annually. ([runway.com](https://runway.com/pricing?tool=runway&utm_source=openai)) **Google Flow/Veo 3.1** is the quality pick. Its Ingredients-to-Video workflow uses reference images to maintain character, background and object consistency. It supports native vertical output, generated audio and 1080p/4K upscaling. ([blog.google](https://blog.google/innovation-and-ai/technology/ai/veo-3-1-ingredients-to-video/?utm_source=openai)) ## Recommended production workflow 1. **Design one character** with a distinctive silhouette, hairstyle, outfit and color palette. 2. Produce a reference sheet containing: - Front portrait - Three-quarter view - Profile - Full-body view - Neutral expression 3. Train/save the identity in **Higgsfield Soul ID**. 4. Generate individual **5–8 second vertical shots**, requesting only one action per shot. 5. Assemble four to seven shots into a **20–35 second TikTok**. 6. Add narration, captions, sound effects and music during editing rather than asking the generator to create the entire finished TikTok in one pass. 7. Reuse the same identity asset, wardrobe description, reference images and voice settings for every episode. **Important limitation:** no generative-video platform guarantees perfect consistency. Hands, clothing details and facial features can still drift, especially during fast movement, extreme angles or major wardrobe changes. Short shots and strong reference images improve the success rate. ## Bottom line - **Choose Higgsfield** for the easiest recurring-character TikTok workflow. - **Choose Runway** if cinematic control and scene building matter
Sources: geo.higgsfield.ai · runwayml.com · runway.com · blog.google
chatgpt.com, logged out, no account, default model, web search on by default; conversation https://chatgpt.com/uc/6a9b66fd-a3bc-83ea-a263-249807534414
If your #1 requirement is consistent recurring characters for a faceless TikTok channel, I'd rank them: Rank | Tool | Character consistency | Best for 1 | Runway Gen-4/4.5 | 5/5 | Recurring characters + cinematic storytelling 2 | Google Veo 3.1 | 5/5 | High-quality realistic/animated clips 3 | Kling | 4.5/5 | Character animation + action 4 | HeyGen | 4/5 | Talking/avatar-style characters 5 | InVideo AI | 3/5 | Fast automated TikTok production My pick: Runway For your use case, I'd choose Runway. Its Gen-4 References system is specifically designed to maintain a character's appearance across different scenes, lighting, locations and treatments. You can save a character reference and reuse it, and Runway supports up to three references in a generation. That makes it particularly good for a channel like: "A sarcastic raccoon explains weird historical events". You could establish one raccoon character once, then generate: Raccoon in Ancient Rome, Raccoon at a medieval castle, Raccoon discovering electricity, Raccoon interviewing historical figures, Raccoon reacting to bizarre facts — while keeping the same character identity and visual style. Runway also has workflows specifically aimed at going from a single character image to a short film and maintaining consistent characters/scenes. Veo 3.1 Google DeepMind Veo is another excellent choice. Veo 3.1 explicitly supports reference images for characters and scenes, and Google demonstrates keeping the same character across different environments. It also generates audio. I'd lean toward Veo if your priority is maximum visual realism, while Runway is my preference for a repeatable character-driven content workflow. What I'd actually use for a TikTok channel I'd build the pipeline like this: Character design → Runway → voice → CapCut → TikTok Create a master character image. Create 5–10 additional reference poses/expressions. Save the character as a reusable reference in Runway. Generate individual 5–10 second shots rather than trying to make the entire TikTok in one generation. Use the same character reference for every shot. Add narration with an AI voice. Assemble everything in CapCut, with aggressive captions, sound effects and music. Export 9:16. Important: Don't generate your character from scratch every time. The biggest consistency improvement comes from treating your character image as a persistent asset and feeding it back into every generation. If you tell me what kind of faceless TikTok you want to make (e.g. cartoon stories, horror, history, Reddit stories, motivation, animals, AI influencer, finance), I can give you a specific tool stack + workflow + prompts for producing 30–60 videos/month.
Sources: runwayml.com · deepmind.google
AI video generator with built-in captions and AI voiceover
API named: InVideo, Synthesia, VEED, Faceless.video
Consumer named: CapCut, Descript, VEED
Panel, OpenAI through OpenRouter, native retrieval
## Best pick: **VEED** VEED is the strongest all-in-one choice for your requirements. It can generate a video from a prompt, script, or image and include: - AI-generated visuals - AI voiceover or voice cloning - Automatic subtitles - Custom caption styles - Aspect-ratio presets for TikTok, Reels, and YouTube - A full browser-based editor for correcting narration, captions, timing, and footage ([veed.io](https://www.veed.io/tools/ai-video)) **Limitation:** Generative video models use AI credits, with different models consuming different amounts. The free allowance is limited. ([veed.io](https://www.veed.io/tools/ai-video)) ### Other good options **InVideo AI — best for faceless social and YouTube videos** Creates a complete video from a prompt with stock or generated footage, synchronized AI narration, and animated subtitles. The Plus plan is listed at **$17/month when billed annually**, with 750 monthly credits and watermark-free exports. Credits don’t roll over, and generation costs vary by model. ([info.invideo.io](https://info.invideo.io/text-to-video-generator-adds-subtitles-voiceovers?utm_source=openai)) **Synthesia — best for training and presenter videos** Ideal when you want an AI avatar speaking your script. It supports AI voiceovers in 160+ languages, automatic closed captions, and customizable dynamic captions. Starter costs **$29 month-to-month or $18/month billed annually**, with 10 video minutes per month. It is less suited to fast-paced, footage-heavy social videos. ([synthesia.io](https://www.synthesia.io/features/ai-video-generator?utm_source=openai)) ### Quick decision - **Social media and flexible editing:** VEED - **Automated faceless videos:** InVideo AI - **AI presenter, courses, or corporate training:** Synthesia
Sources: veed.io · info.invideo.io · synthesia.io
chatgpt.com, logged out, no account, default model, web search on by default; conversation https://chatgpt.com/uc/6a9b6754-5c64-83ea-bb71-e585c84250e2
If you mean a tool that can take a script/prompt and automatically create the video, add captions, and generate an AI voiceover, these are strong options: CapCut — Probably the easiest all-in-one choice. It offers AI video generation, AI text-to-speech/voiceover, and automatic captions. VEED — Excellent for social-media videos. Its AI text-to-video workflow can generate visuals, narration, and automatic subtitles from a prompt or script. Descript — Best if you want more control over editing. It has AI-generated video, realistic AI voices/voice cloning, and instant captions. My pick For TikTok/Reels/YouTube Shorts → CapCut. For turning a written script into a polished narrated video → VEED. For professional editing + AI voice → Descript. If you tell me your budget (free/paid) and whether you want realistic AI footage, stock footage, or an AI avatar, I can narrow it down to the best 2–3 options.
Sources: capcut.com · veed.io · descript.com
Frutti AI alternatives for making brainrot videos
API named: Revid, CapCut, Frutti, StoryShort, Hedra, Faceless.video, Viggle, FlowShorts, Media.io
Consumer named: Revid, Pika, Veo, Zebracat, Frutti, StoryShort, Hedra, Clipfly, Eliro
Panel, OpenAI through OpenRouter, native retrieval
Frutti’s main advantage is that it generates a **complete character-driven episode**—story, recurring cast, dialogue, voices, animation and captions—from one premise. Many alternatives specialize instead in gameplay-background brainrot or individual animated shots. ([frutti.ai](https://frutti.ai/generators?utm_source=openai)) ## Best Frutti AI alternatives | Tool | Best for | Starting cost* | Main limitation | |---|---|---:|---| | **StoryShort** | Closest all-in-one alternative | From $23/mo annually | Less control over individual scenes | | **Revid AI** | Fast gameplay-style brainrot | Free to try | More template-driven than Frutti | | **Brainrotify** | Cheapest dedicated option | Free; Pro $7.99/mo | Not as strong for recurring character series | | **FlowShorts** | Automated channel posting | Free first video; $19/mo | Geared toward faceless videos rather than character dramas | | **Media.io** | Italian-brainrot creatures and surreal clips | Free credits | Often produces clips/assets rather than polished episodes | | **Hedra** | Talking characters and dialogue performance | Limited free use | Generally one character shot at a time | | **Viggle** | Free character animation and meme motions | 5 free videos/day | No complete script-to-episode pipeline | | **CapCut** | Maximum editing control on a budget | Free option | More manual work | \*Pricing and allowances checked September 2026 and can change. ### 1. StoryShort — best overall substitute StoryShort can turn prompts, scripts, stories or PDFs into videos with narration, gameplay backgrounds, captions and music. It supports custom/consistent characters, voice cloning and several aspect ratios. Its listed Starter plan is $23/month when billed annually, with roughly 25–40 short videos. ([storyshort.ai](https://storyshort.ai/tools/text-to-brainrot?utm_source=openai)) **Choose it if:** you want the closest thing to entering an idea and receiving a finished video. **Limitation:** Its gameplay/story workflow is closer to conventional faceless brainrot than Frutti’s fully performed ensemble episodes. ### 2. Revid AI — best for Subway Surfers/Minecraft brainrot Revid takes an idea or script and automatically adds an AI voice, animated captions, music and gameplay footage. It supports split-screen formats and browser-based generation, with a free trial requiring no card. ([revid.ai](https://www.revid.ai/tools/text-to-brainrot?utm_source=openai)) **Choose it if:** your target format is dramatic narration over Minecraft, GTA or Subway Surfers. **Limitation:** It is optimized for fast, repeatable short-form content—not persistent characters and original animated story worlds. ### 3. Brainrotify — best budget choice Brainrotify supports brainrot, stories, Reddit posts, iMessage conversations, quizzes and other templates. Its free tier provides daily credits with no export watermark; Pro is listed at $7.99/month. It also offers batch generation and an API. ([brainrotify.me](https://www.brainrotify.me/Pricing?utm_source=openai)) **Choose it if:** you want inexpensive, high-volume testing or automation. **Limitation:** The standard outputs are more template-based than Frutti’s generated character dramas. ### 4. FlowShorts — best for automated posting FlowShorts generates scripts, visuals, narration and captions, then can automatically publish to TikTok, YouTube and Instagram. Plans currently include one free video, $19/month for eight videos, and $39/month for 30. ([flowshorts.app](https://flowshorts.app/?utm_source=openai)) **Choose it if:** running an automated faceless channel matters more than custom character animation. **Limitation:** It is an automation platform first; it’s not the closest match for talking-fruit or recurring-cast episodes. ### 5. Media.io — best for creating original Italian-brainrot creatures Media.io supports text-to-image, image-to-image and image-to-video generation for surreal characters and memes. You can create the character, remix an existing image and animate it. ([media.io](https://www.media.io/italian-brainrot-generator-ai.html?utm_source=openai)) **Choose it if:** you mainly need strange creatures, transformations and short surreal scenes. **Limitation:** Expect to add dialogue, captions and episode structure separately in an editor. ### 6. Hedra — best for talking-character performances Hedra’s Character model creates audio-driven talking-character videos with
Sources: frutti.ai · storyshort.ai · revid.ai · brainrotify.me · flowshorts.app · media.io
chatgpt.com, logged out, no account, default model, web search on by default; conversation https://chatgpt.com/uc/6a9b6784-6690-83ea-bd18-8f73b1614718
If you mean Frutti AI-style brainrot videos—absurd characters, AI voices, captions, fast pacing, and 9:16 Shorts/Reels—these are the alternatives I'd look at: StoryShort — probably the closest if you want the classic Subway Surfers/gameplay + AI narration + captions format. Fast and relatively low-effort. Revid AI — good for generating short narrated videos and has publishing/automation features. Better for faceless content than recurring-character stories. Zebracat — one-prompt workflow that can assemble narration, visuals, music/SFX and captions. Good if you want something close to "prompt → finished short." Hedra — especially good if you have a specific brainrot character/image and want to make it talk and lip-sync. You handle the script/audio yourself. Veo — better if you want higher-quality original AI-generated scenes rather than the usual gameplay-background format. More manual work, but much more visual freedom. Pika — useful for weird/meme animations and compositing; a good choice if you want more control instead of a fully automated brainrot pipeline. Clipfly — useful for trying brainrot formats cheaply and then editing the generated clip yourself. Eliro — specifically targets the classic brainrot formula with split-screen gameplay, AI voiceover, animated captions and story/Reddit templates. My picks For easiest Frutti replacement: → Zebracat For Subway Surfers/Minecraft-style brainrot: → StoryShort or Eliro For talking AI creatures: → Hedra For genuinely crazy original visuals: → Veo For maximum automation/volume: → Revid AI One important distinction: Frutti is unusual because it tries to produce a complete episode with recurring characters, dialogue, voices, captions and vertical export in one workflow. Most alternatives give you only part of that pipeline. If you tell me whether you want free, best quality, or "one prompt → finished TikTok", I can narrow it down to the best 3.
Sources: frutti.ai · openaitoolshub · eliro
Method, and what it cannot claim
- Six pairs. Enough to publish, not enough to generalise. The study is designed to be repeated; the harness and lexicon are in the repository.
- One consumer surface. Claude.ai and Gemini require an account, and a signed-in account carries memory and personalisation — the exact confound this company was founded to separate. Their API-side answers were recorded and are in the raw data, unpaired.
- Brands are counted against a fixed lexicon built from both sides' answers before comparison. A brand not in it is invisible to every rate above. Brand names that are also ordinary words ("Captions") are matched only through unambiguous forms.
- Cited-domain overlap is low confidence. Logged-out ChatGPT shows sources as labelled chips, not links; labels were mapped to domains where unambiguous.
- One pass, one day, temperature 0.2 on the panel side, default on the consumer side. A single run is noise; the same is true here.
- Panel cost: $1.78 for eighteen answers with native retrieval, two OpenAI timeouts retried, one empty $0.21 answer retried and now treated as an error by the engine.
Raw data
Panel answers (18, three families) · Consumer answers (6) · Computed comparison · Brand lexicon · Questions
Why publish a number this size? Because the alternative is to keep saying "APIs may differ" without ever finding out, and a company that sells measurement should be the first to measure its own instrument. A second pass, on more questions, is the next entry in the observation log.