https://reloop.so

You are helping me make my product visible and recommendable by AI systems — ChatGPT, Claude, Gemini, and Perplexity.

## My Product

**Name:** Reloop
**Website:** https://reloop.so
**What it does:** Create scroll-stopping AI video ads with realistic avatars. Generate UGC-style AI ads in minutes — no scripts, no editing, no prompts. 200 free credits.

## AI Visibility Audit Results

**Current AI Exposure Score: 68/100**
> Your score of 68/100 is above average. Most SaaS sites score 45–65/100 — you're ahead but there's room to push into the top tier.
**Projected score after all fixes: 98/100**

### Score by category
- AI Crawl Access: 19/27
- Content Quality: 13/27
- Product Clarity: 15/15
- Structured Data & Meta: 27/33
- Agent Readiness: 3/10
- Trust & Social Proof: 10/15
- EEAT & Discoverability: 12/18

### AI Crawler Access

- NOT EXPLICITLY ALLOWED: ChatGPT, Claude, Perplexity, Gemini, Meta AI, Apple AI — add explicit Allow rules

## Failing Checks (grouped by effort)

### Phase 1 — Easy wins (do these first)
- [AI Crawl Access] llms.txt exists (5 min): No llms.txt found. This file tells AI systems exactly what your product is. We generated one for you below.
- [AI Crawl Access] llm.json exists (5 min): No llm.json found. This machine-readable JSON file lets AI agents programmatically access your product name, features, pricing, and integrations. We generated one for you below.
- [AI Crawl Access] llms-full.txt exists (30 min): No llms-full.txt found. This is the expanded version of llms.txt — a single Markdown file containing your full product documentation, feature details, use cases, and pricing. Long-context models like Gemini 1.5 Pro prefer this over crawling individual pages.
- [Content Quality] Data density (tables & lists) (20 min): Low data density: 0 table(s) and 0 list(s). AI models extract facts from <table> and <ul> tags much more reliably than from prose. Your pricing, features, and comparisons should use tables and bullet points — not paragraphs.
- [Structured Data & Meta] Organization schema (5 min): No Organization schema. This tells AI systems your company name, logo, and social profiles — critical for accurate brand identification.
- [Structured Data & Meta] Schema sameAs entity verification (10 min): No sameAs property in your JSON-LD schema. AI systems triangulate your brand — they check if your site, LinkedIn, Twitter/X, ProductHunt, and GitHub all describe you consistently. Without sameAs links, AI has lower confidence in recommending you.
- [Trust & Social Proof] Quantifiable metrics (15 min): No quantifiable metrics found (e.g. '10,000+ users', '500+ teams'). Specific numbers help AI recommend with confidence.
- [EEAT & Discoverability] Social profiles linked (10 min): No social profile links found. EEAT Authoritativeness — linking to your Twitter/X, LinkedIn, and GitHub helps AI systems verify your brand identity and find corroborating information about your product.

### Phase 2 — Medium effort
- [Content Quality] Text-to-HTML ratio (1 hour): Low text-to-HTML ratio (4%). Your site may rely heavily on JavaScript rendering. AI crawlers often get empty content from JS-heavy sites.
- [Content Quality] Answer-first content structure (30 min): Low answer-first score (0%). Your H2 sections begin with marketing fluff instead of direct answers. ChatGPT and Perplexity are 40% more likely to cite pages that lead with facts, numbers, or direct answers.
- [Agent Readiness] Integrations clarity (1 hour): No integrations info found. AI cannot determine what your product connects to.
- [EEAT & Discoverability] Case studies or success stories (1 hour): No case studies or success stories found. EEAT Experience — showing real results ("Company X increased Y by Z%") is the strongest signal that you have first-hand experience delivering value.

### Phase 3 — Bigger investments
- [Agent Readiness] Documentation or API reference (days): No docs or API page found. AI agents cannot assess technical capabilities.

### Missing product information
- No documentation or API reference found
- No integrations information found

### Full report
https://www.aiexposuretool.com/stats/reloop-so-q7rjes

## What I need from you

Fix every issue above. Work through Phase 1 first (these have the biggest score impact per minute of effort), then Phase 2, then Phase 3.

For each fix give me:
1. **Exact code or copy** — no placeholders, no "add your text here". Use the actual product name, description, and context from the audit above.
2. **Where to put it** — file name, line, or section
3. **Which AI systems this helps** and why

### Specific outputs I need

1. **robots.txt** — the complete file content that explicitly allows GPTBot, ClaudeBot, PerplexityBot, Google-Extended, Meta-ExternalAgent, OAI-SearchBot, and Applebot

2. **llms.txt** — a complete, accurate llms.txt file (200–400 words) for this product. Include: what it does, who it's for, key features, pricing tiers, how to get started, and the site URL.

3. **llm.json** — a complete JSON file with: name, url, description, category, pricing (array of plans), features (array), integrations (array), target_audience (array)

4. **JSON-LD structured data** — a complete `<script type="application/ld+json">` block with SoftwareApplication schema, plus a separate Organization schema with sameAs links to social profiles

5. **FAQPage schema** — 6–8 Q&A pairs as a FAQPage JSON-LD block covering: what the product does, who it's for, pricing, how to get started, what makes it different

6. **Homepage meta tags** — exact HTML for `<title>`, `<meta name="description">`, og:title, og:description, og:image, og:url, og:type, and `<link rel="canonical">`

7. **H1 and subheadline rewrite** — new homepage H1 and subheadline paragraph. Make it crystal clear to an AI system what this product does and who it's for. Show before → after.

8. **About page outline** — a brief /about page outline with founder/team context, founding story, and contact info (EEAT Experience signal)

9. **Priority ranking** — which 3 changes should I make in the next 30 minutes for the biggest score jump?

All output must be copy-paste ready. No vague suggestions.