LunarCrush provides real-time insight into the cryptocurrency, stock, and NFT marketplaces. The platform decodes social sentiment and market trends, tracking volume and engagement metrics for over 20,000 assets across a broad selection of digital platforms. Rich analytics features like social metrics dashboards, qualitative and quantitative sentiment analysis, and custom alerts expose critical data to simplify market analysis and enhance trading strategies. Evolving rapidly since its 2018 launch, LunarCrush leverages rapidly developing AI and machine learning technology to deliver market intelligence. Responding to growing interest in the cryptocurrency marketplace, the platform has transformed from a niche data and social media aggregator into a full-scale analytics platform. The customer sought to enhance application security and availability while simplifying infrastructure to reduce backend complexity. They aimed to eliminate surprise cloud costs, avoid vendor lock-in, and seamlessly scale analytics and media features with serverless compute and edge storage. Additionally, they prioritized secure developer access to tools without relying on public exposure or manual VPN configurations.
LunarCrush’s frustration with traditional cloud hosting services began when they launched their mobile app, incentivizing new memberships with giveaways of LUNR, their promotional crypto token. International interest pushed the limits of their cloud infrastructure, as millions of new signups and hundreds of thousands of concurrent transactions bombarded the app. The company’s AWS services — EC2 and Aurora — handled the load, but at a cost. “Working with major cloud providers came with some hard lessons. By committing to a long-term agreement to limit expenses, we locked in our backend architecture costs,” explains Dan Williams, Chief Technical Officer at LunarCrush. “As we evolved the platform, improved our efficiency, and streamlined our business model, we were paying for computing power and bandwidth we didn’t need.”
As LunarCrush shifted their offering from a high-volume free service to a lower-traffic subscription model, they were shouldering AWS service fees exceeding $25,000. A new infrastructure strategy was needed.They began redesigning their cloud architecture and used Cloudflare to offload performance, security, and edge computing to a single, integrated platform.
“Our goal was to make LunarCrush as fast and secure as possible while avoiding the 'gotchas' — costs most cloud providers don’t tell you about,” says Wiliams. “Using Cloudflare serverless computing and cloud storage solutions as the public gateway to our private infrastructure was the best way to get there.”
The journey to retake control of their architecture spanned multiple phases. LunarCrush began by implementing Cloudflare’s global CDN to accelerate static asset and image delivery while eliminating unnecessary roundtrips to their origin servers and storage. Adding another AWS service was not an option.
“Delivering content with AWS involved turning on, managing, and paying for multiple services,” says Williams. “Cloudflare offered a centralized alternative to expensive load balancers and provided additional equipment, free bandwidth, and a web interface that was far better than any other offering we had seen.”
Using Cloudflare, LunarCrush also secured its public applications, accelerated service delivery, and improved their global user experience. Global threat intelligence and Cloudflare application security services protect the company’s public web properties and backend servers from traffic spikes and automated attacks — without adding overhead.
“We've experienced several DDoS and malicious bot attacks since moving to Cloudflare, but they never make it as far as our origins. I don’t think there is a better solution for web security,” says Williams. “Now and then, somebody writes some code trying to scrape us or hit our API, but Cloudflare makes us a difficult target.”
LunarCrush rounded out their first-phase transformation with the Cloudflare developer platform, Workers and Durable Objects — dynamic serverless computing and state storage functionality on Cloudflare global serverless infrastructure — helped secure and streamline their frontend.
“We began using Workers as soon as Cloudflare introduced them,” says Williams. “The first iteration with Workers was about setting up the logic to protect our origin and cache. This verified that an authorization token was valid and used Durable Objects to keep that in memory for quick lookups and ensure fast execution. Leveraging waitUntil, we buffered logs in Durable Objects after delivering a response, so we missed no requests when writing to our backend.”
Using Workers and Pages on the unified Cloudflare platform, LunarCrush retired Vercel, the front-end cloud infrastructure framework powering their website and first-generation mobile application. Cloudflare's transparent, predictable pricing and flexible service agreements addressed LunarCrush’s budgeting challenges while providing the company with the ability to distinguish between human and automated traffic.
“When our application hit hyper-scale, we incurred massive unexpected costs for infrastructure usage that did not correlate with revenue,” says Williams, “Cloudflare was the only replacement solution to provide the observability and pricing necessary to handle this scenario, helping us identify and mitigate traffic spikes before they could affect the platform or our bottom line.”
The switch yielded significant savings.
“Before we moved to Cloudflare, we were seeing unexpected invoices that reached $100,000,” says Williams. “Cloudflare is very predictable — now our biggest surprise is an extra $500, maybe $1000, numbers we can manage.”
The company has steadily expanded their use of Cloudflare Workers and Cloudflare’s developer platform to power high-frequency user-facing services. This includes what Williams calls “user-scale” services like developer endpoints, dynamic images, and live data overlays. These overlays — the heart of the platform — display trending content and metrics across LunarCrush’s livestreams, dashboards, and analytics tools.
“Our top priorities are making sure the site always works and we can handle everything coming our way,” says Williams. “We depend 100% on Cloudflare to handle user scale and ensure our website is instantly fast so we don't have to worry about it.” Offloading live components to Cloudflare, LunarCrush reduces back-end server and computing loads and ensures a consistent global user experience, irrespective of volume, market volatility, or influencer-driven traffic increases.
Using Workers AI to run AI applications without the burden of maintaining or paying for unused infrastructure, LunarCrush powers What’s Up, an AI companion that summarizes key discussions from millions of social media posts. The company also relies on Workers AI to dynamically generate and regularly refresh fallback images for trending conversations that lack defining visual assets. And, with Cloudflare R2 — a zero-egress fee and S3-compatible object storage on a global network — they can easily host AI-generated visuals and user-uploaded avatars.
“All our media content and CDN images are in R2,” says Williams. “Our infrastructure doesn’t rely on any cloud providers anymore — it's all Cloudflare and some machines we use to process data in the background.”
Consolidating delivery, logic, and storage on the Cloudflare network, LunarCrush also avoids the overhead of scaling backend systems to keep pace with demand. Williams estimates that by moving their most traffic-intensive workloads to Cloudflare, this has reduced LunarCrush’s monthly infrastructure fees from $25,000 to $5,000 — a savings of 80%.
Adopting Cloudflare Zero Trust Network Access has helped the company simplify developer workflows and secure and streamline access to cloud-hosted production environments and internal tools. Previously, LunarCrush developers relied on publicly available bastion hosts within AWS to access staging environments — an arrangement that slowed development and exposed infrastructure to unnecessary risk.
“Zero Trust was a big deal for us, allowing our developers to collaborate securely,” says Williams. “Using Cloudflare, you can host from your laptop, prototype whatever you want, and it’s already set up if you need to move forward — all without public exposure or deploying to production.”
The change significantly improved developer velocity and flexibility — all protected on the Cloudflare network.
Although LunarCrush initially adopted Cloudflare Tunnel to streamline developer access and eliminate the overhead of bastion hosts, the Zero Trust solution has helped the company address another costly infrastructure issue — inter-regional data transfer fees. “We were paying significant fees just to keep multiple regions in sync — something we were unaware of until we investigated our costs,” says Williams. “Using a combination of Tunnel and Workers to route traffic directly through Cloudflare, we avoided the need to replicate backend infrastructure for redundancy.”
Using the Cloudflare connectivity cloud — a secure, unified platform of cloud-native services — gave LunarCrush the flexibility to replatform backend services without the pitfalls of vendor lock-in. William describes an incident where he discovered their previous provider had disabled the migrate command to hinder data portability in a proprietary version of a popular open source data platform.
“They modified the software to make it harder to migrate out,” he explains.
Decoupling application delivery and security from their underlying cloud infrastructure with Cloudflare, LunarCrush bypassed these constraints, keeping their public-facing applications stable and responsive.
“As our needs evolved, we moved our backend to different cloud providers several times,” says Williams. “Cloudflare is the key to being provider-agnostic and completing those zero-downtime migrations.”
LunarCrush intends to continue evolving their architecture to take greater advantage of Cloudflare services, for example, exploring ways to sync platform data using R2 and Workers to eliminate dedicated database instances. Williams also monitors new Cloudflare product releases for features that can extend platform capabilities and improve performance, security, and cost efficiency.
“I often go through the Cloudflare developer blog to see what new products are coming out and what state they're in,” says Williams. “Based on Cloudflare’s reputation and our history of success, if there’s something viable in the pipeline, we will adopt it.”
Reduced cloud infrastructure costs by 80% — down to US $5000 from US $25,000 monthly
Mitigates repeated DDoS and bot attacks, protecting and isolating origin servers on the global network
Replaced public-facing bastion hosts and accelerated developer workflows with secure, easy-to-maintain policy-based Zero Trust access
Eliminated surprise inter-regional data transfer charges and cloud platform lock-in
Enhanced dynamic media generation offloading compute and storage intensive tasks and AI inference to Workers AI and R2
“We've experienced several DDoS and malicious bot attacks since moving to Cloudflare, but they never make it as far as our origins. I don’t think there is a better solution for web security.”
Dan Williams
Chief Technical Officer, LunarCrush
“Before we changed to Cloudflare, we were seeing unexpected invoices that reached US $100,000. Because Cloudflare is very predictable, now our biggest surprises are an extra US $500, maybe $1000 — those are numbers we can manage.”
Dan Williams
Chief Technical Officer, LunarCrush