modular.com
success
Claude Code
intent

What does modular.com do and who is it for? Explain it back to me.

18steps
62.7sduration
$0.9229cost
235,454tokens
21 steps3 reasoning steps4 searches
home
/max
docs
/cloud
docs
search
search
/inference/dedicated-e…
/batch-inference
/inference/shared-endp…
search
/platform
/mojo
/company/about
search
/products
/hosted
docs
79%
on-site discovery
64%
reliability
57%
link following
path origin
  • previous resource57%
  • web search21%
  • prior knowledge21%
insight

The agent successfully understood Modular's business model, products, pricing tiers, and competitive differentiation by fetching the homepage, product pages (MAX, Mojo), pricing page, company about page, and cloud service endpoint pages. The site was reasonably navigable for product discovery but lacked transparency on specific pricing numbers and had broken URLs (/products, /platform, /cloud, /batch-inference all returned 404), forcing the agent to assemble answers from fragments and rely on web search to fill gaps.

  • ›The homepage [1], MAX page [2], and Mojo page [7] provided clear product positioning; the dedicated endpoints [16] and shared endpoints [18] pages exposed the cloud service offerings with service-level differentiation but without published per-token or per-minute rates.
  • ›The pricing page [3] stated 'Free to deploy and manage yourself' and referenced contact-based enterprise pricing, but did not surface concrete pricing numbers—the agent had to infer the model from endpoint page descriptions and use web search [15] to discover the pricing page URLs themselves.
  • ›Multiple expected URLs returned 404 (/products, /platform, /cloud, /batch-inference, /hosted), indicating poor information architecture for agent navigation; the agent compensated by using prior knowledge and web search to locate /inference/shared-endpoints and /inference/dedicated-endpoints.
  • ›The about page [8] and blog content (referenced in search results) provided company context and Qualcomm acquisition news but did not explain how this affects product roadmap or long-term positioning.
  • ›The site did not publish benchmark details or detailed per-token/per-minute pricing—both critical for evaluating cost claims ('2x performance', '85% cheaper') and comparing against vLLM, TensorRT-LLM, and SGLang.

See every agent that visits modular.com

Front builds the agent version of your site and shows which agents visit, what they came for, and what your pages couldn't answer.

or talk to us about agent readiness →