Serverless is a computing model where cloud providers manage server infrastructure, billing only for code execution time.
Serverless (AWS Lambda, Cloudflare Workers, Vercel Functions, Google Cloud Run) automatically scales 0 to thousands of concurrent requests. Pay-per-request model rewards efficient code. Use cases: APIs, scheduled jobs, event-driven workflows. Drawbacks: cold starts, vendor lock-in, debugging complexity. By 2026, "serverless-first" is the default for greenfield projects.
Serverless removes capacity planning for variable workloads and shifts pricing from "what you provisioned" to "what you used." For event-driven and bursty workloads, it can be dramatically cheaper than dedicated capacity.
A webhook handler runs as a single function on AWS Lambda. When events arrive, it spins up, processes them and shuts down; idle hours cost nothing, while peak hours scale automatically without any capacity planning.
Serverless is not literally without servers. The provider runs and scales them for you, which makes them feel invisible but does not change the underlying mechanics.
Be wary of cold starts on latency-sensitive paths; either keep the function warm or use provisioned concurrency for user-facing endpoints.
Serverless falls under the Hosting category.
These tools put serverless into practice. Compare features, pricing, and ratings:
Now that you understand Serverless, explore the best tools in this category.