runwayml.com — AI Site Grade

Analysis

Runwayml.com: AI-Visibility Audit

The site's most consequential AI-visibility gap is not a block — it is the absence of an llms.txt file, which means every AI crawler that lands on the homepage gets a 746KB Next.js bundle instead of a structured content map, despite all bots receiving full 200 access.

Crawler Access

All eleven tested AI crawlers — GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot, ChatGPT-User, Bytespider, Applebot-Extended, anthropic-ai, Perplexity-User, and a baseline browser — receive a 200 status with identical 746KB payload from Vercel's edge. No UA-based blocking, no Cloudflare challenge, no 403. The robots.txt is a single Allow: / rule for User-agent: * with zero AI-bot-specific directives. This is permissive but naive: every crawler gets the full JS-heavy page, not a distilled AI-friendly version. The llms.txt returns a 404 (Next.js error page), forfeiting the emerging standard for declaring an AI content map.

Cold-Knowledge Gap

The LLM's prior knowledge of Runway is stale by roughly 18 months. It knows about Gen-1 and Gen-2 (text-to-video, video-to-video) and mentions a $50M Google investment and training-data copyright controversy. The actual site has moved far beyond: Gen-4.5 (launched November 2025) is the flagship, alongside GWM-1 (December 2025) — a General World Model family with three variants (Worlds, Avatars, Robotics). The cold model knows nothing about GWM-1, Runway Characters (real-time video agents), Aleph 2.0 (video editing), Seedance 2.0, or the NVIDIA/Lionsgate partnerships. The copyright controversy that the model recalls is entirely absent from the site — no mention, no rebuttal, no trust page.

Schema Posture

The homepage and product page carry rich Organization, SoftwareApplication, and WebSite JSON-LD with founders, founding date, social profiles, and contact info. The pricing page has Offer schema for all five plans (Free at $0, Standard at $12/mo, Pro at $28/mo, Unlimited at $76/mo, Enterprise). Research articles use TechArticle with datePublished and BreadcrumbList. However, the /news page returns zero visible text (a JS shell) and no schema at all — any AI crawler indexing that page gets an empty content signal. The /about and /research landing pages also lack schema entirely.

External Signals

DNS TXT records confirm domain verification for Anthropic, Perplexity, Apple, Google, Slack, Stripe, Zoom, and Jamf — indicating active integration with or verification by multiple AI platforms. The site prominently features partnerships with NVIDIA (Vera Rubin architecture collaboration), Lionsgate (film production), UCLA film school, and KPF (architecture). The AI Film Festival (AIFF) and Gen48 event subdomains signal community engagement. No external press or review results surfaced in search — the brand's external narrative is almost entirely self-published.

Surprising Findings

The /news page is a Next.js shell with 0 words of visible text — a critical blind spot for any AI crawler trying to index company announcements. The sitemap.xml contains only 29 URLs, which is thin for a company with multiple product lines, research publications, customer stories, and app pages. Several URLs referenced in navigation (e.g., /solutions, /studios, /enterprise-contact) are absent from the sitemap. The dateModified on the API page schema reads 2026-05-29 — a future date that may confuse temporal reasoning in knowledge graphs.

Findings

1. Missing llms.txt file blocks AI content map High

   The site returns a 404 for /llms.txt, forfeiting the emerging standard for declaring an AI-friendly content map. All AI crawlers receive a 746KB Next.js bundle instead of structured content.

   What to change: Create an llms.txt file at the root that lists key pages (product, pricing, research, about) and provides a brief site summary for AI crawlers.

2. News page renders as empty JavaScript shell High

   The /news page returns a 200 status but contains zero visible text — a Next.js shell with no content for AI crawlers. Any bot indexing this page receives an empty content signal.

   What to change: Ensure the /news page renders static HTML content or implement server-side rendering so crawlers can read article titles and summaries.

3. LLM cold knowledge is 18 months behind current products High

   The LLM's prior knowledge of Runway is limited to Gen-1 and Gen-2, a $50M Google investment, and a training-data copyright controversy. It knows nothing about Gen-4.5, GWM-1, Runway Characters, Aleph 2.0, Seedance 2.0, or partnerships with NVIDIA and Lionsgate.

   What to change: Publish dedicated pages for each major product and model (Gen-4.5, GWM-1, Characters) with clear, crawlable descriptions and schema markup to accelerate knowledge refresh.

4. Sitemap contains only 29 URLs, missing key pages Medium

   The sitemap.xml lists only 29 URLs, which is thin for a company with multiple product lines, research publications, and customer stories. Several navigation pages like /solutions, /studios, and /enterprise-contact are absent.

   What to change: Expand the sitemap to include all important pages, especially product subpages, research articles, and customer stories.

5. API page schema has a future dateModified (2026-05-29) Medium

   The JSON-LD on the /api page includes a dateModified of 2026-05-29, which may confuse temporal reasoning in knowledge graphs and reduce trustworthiness.

   What to change: Update the dateModified field to the actual last modification date or remove it if not accurately maintained.

6. About and research landing pages lack schema markup Medium

   The /about and /research landing pages have no JSON-LD schema, missing an opportunity to provide structured entity data to AI crawlers.

   What to change: Add Organization or WebSite schema to /about and CollectionPage or TechArticle schema to /research landing page.

7. Robots.txt has no AI-bot-specific directives Low

   The robots.txt contains a single Allow: / rule for User-agent: * with no specific directives for AI crawlers like GPTBot or ClaudeBot. While permissive, it misses the chance to guide bots to AI-friendly content.

   What to change: Consider adding specific directives for AI crawlers, e.g., allowing access to key pages and disallowing low-value paths.

8. Training-data copyright controversy absent from site Medium

   The LLM recalls a training-data copyright controversy, but the site contains no mention, rebuttal, or trust page addressing it. This leaves AI crawlers without a counter-narrative.

   What to change: Publish a trust or ethics page that transparently addresses data sourcing and copyright practices.

What's working

• All 11 tested AI crawlers receive full 200 access — Every tested AI bot gets a 200 status with no blocking, challenges, or UA-based restrictions, ensuring the site is fully crawlable.
• Homepage, product, and pricing pages have rich JSON-LD schema — Organization, SoftwareApplication, WebSite, and Offer schemas are present with detailed information including founders, founding date, social profiles, and pricing tiers.
• Research articles use TechArticle schema with dates — Pages like /research/introducing-runway-gen-4.5 and /research/introducing-runway-gwm-1 include TechArticle schema with datePublished and BreadcrumbList, aiding AI understanding.
• DNS TXT records confirm verification by Anthropic, Perplexity, Apple, Google, and more — Domain verification TXT records for multiple AI platforms indicate active integration and trust signals.
• Site prominently features partnerships with NVIDIA, Lionsgate, and others — Partnerships with NVIDIA, Lionsgate, UCLA film school, and KPF are highlighted, providing strong external authority signals.
• AI Film Festival and Gen48 event subdomains signal community engagement — Subdomains for AIFF and Gen48 indicate active community events that can generate external backlinks and mentions.