Vercel · Core platform
A WhatsApp bot waits on the model. An eligibility check waits on a credit bureau. A form waits on Salesforce. On a fixed server you pay for the whole box all month, busy or quiet. Fluid Compute runs many requests on one instance and bills active CPU, provisioned memory and invocations. While a request waits, the instance takes the next one.
Salesforce consulting partner since 2010 · MVP Hall of Fame · 200+ agents in production · UAE, India and US desks
The problem
What Fluid Compute is
Facts from Vercel’s own pages, checked on 16 October 2026. The sources are at the end of this page.
How it works
Five steps. The function runs next to your data, shares its instance while it waits, and finishes the slow writes after the customer has a reply.
Infographic
Step by step
Deploy
Deploy the function
An API route, a server action or a webhook. Fluid is on for new projects.
Region
Run it next to the data
Pick the region nearest your database. Mumbai is the closest to the Gulf.
Share
Requests share warm instances
While one request waits on a model, the same instance takes the next one.
After
Finish the work after the reply
waitUntil writes to Salesforce or a log once the customer has an answer.
Bill
Pay for active CPU
Plus provisioned memory and invocations, priced per region.
How it fits together
Requests arrive from your website, a WhatsApp webhook or an agent. The CDN answers what it can. Fluid functions do the rest: call a model through AI Gateway, read your database, write to Salesforce. Observability shows which call is slow.
Infographic
Use cases · worked examples
How Fluid Compute fits into the work of the industries we serve, one step at a time.
Example 1 · Real estate, Dubai · Infographic
Today: Enquiries come in on WhatsApp at all hours. The bot that answers them runs on a server a freelancer set up, and it goes quiet when a new listing brings a rush.
After: The WhatsApp webhook runs as a Fluid function. It answers the buyer, then writes the lead to Salesforce with waitUntil, after the reply has gone.
Worked example, step by step
Evening enquiries get an answer in the chat and a lead in Salesforce by morning.
Example 2 · Financial services, Mumbai · Infographic
Today: The eligibility check calls a credit bureau and a bank statement service, and each takes several seconds. The first version ran in Washington, D.C., the default region, while the database sits in Mumbai.
After: The function moves to the Mumbai region next to the database. While it waits on the bureau, the instance serves other applicants, and the bill counts active CPU.
Worked example, step by step
The check runs next to the data, and time spent waiting on the bureau stops showing up as CPU.
Example 3 · IT service desks, Austin · Infographic
Today: When a client’s email goes down, hundreds of tickets arrive in minutes. The triage service falls over, and one bad ticket takes the rest down with it.
After: Triage runs on Fluid. Tickets share warm instances, new ones start as the queue grows, and an unhandled error in one ticket leaves the others running.
Worked example, step by step
A morning outage turns into one incident to work, and triage keeps running through it.
Is it for you
Good fit
Not for you yet
What Mindcat does
The assessment is free: a short read of the org and a written note. If you go ahead, the first build week is a fixed AED/SAR 14,900 (about USD 4,050), credited in full when you sign the build within 60 days. After go-live, we keep it patched, watched and changing every month.
FAQ
Sources
Checked against Vercel’s own pages on 16 October 2026. Vercel changes fast, so we check again before any plan we write for you.
Vercel is a product of Vercel Inc. Mindcat builds and runs work on it for clients; we are not affiliated with or endorsed by Vercel.
Start here
One or two lines is enough. You hear back within 24 hours from the person who signs the work.
