Fire & smoke simulation in the cloud

Run FDS at scale.
Manage simulations
then share the results.

Smoke Cloud is the simulation lifecycle platform for fire engineers and fire researchers. Version, share, and analyse every simulation in one place — scale to thousands of cores when an emergency demands it, keep costs under control, and archive runs so they stay searchable for years.

building_1
Running
CREATED
INPUTS
VALIDATED
QUEUED
RUNNING
COMPLETE
08:46:02Uploaded building_1.fds142 MB
08:48:30Validated and allocated meshes
08:50:01Queued simulation
08:51:18Launchedok
compute · n2 · 24 cores
£18.42  / est. £24.10
01 / Platform

Simplify the way simulations are run.

/ 01

Process and workflow

Full revision control over every simulation, with durable storage and clear tracking of inputs, parameters, and runs. Nothing gets lost, nothing gets accidentally overwritten, and every version stays accessible.

/ 02

Share and distribute

Share simulation inputs and outputs with collaborators or clients in a click. No more zipping multi-gigabyte directories, uploading them somewhere, and emailing links — files stay where they are, access stays under your control.

/ 03

Concurrency

When an emergency hits, scale from one run to hundreds in minutes. Smoke Cloud bursts across thousands of cores so a whole team can model every scenario at once — no queue, no waiting room.

/ 04

Budgeting and cost management

Set hard caps per project so simulation costs don't get out of control. Every CPU-hour and byte is logged precisely — ready for onward billing to clients or rolling into next year's forecast.

/ 05

Analysis, review, and auditing

Go beyond raw FDS output with built-in slice viewers, comparison plots, and statistical summaries. Diff two runs side-by-side, replay any historical simulation, and export review-ready reports.

/ 06

Simulation management and archival

Store and index simulations indefinitely, searchable years after the fact — making data retention requirements trivial to satisfy. Lifecycle policies move cold runs to cheaper storage automatically, without losing a single byte.

03 / Interfaces

Run simulations from the terminal, yourcodebase, or any HTTP client.

smoke-cli
For Window, Linux, Mac

Command-line interface

Simply run a simulation via a terminal or script.

# run fire.fds on 8 cores and follow it
$ smoke-cloud run fire.fds --project 12345 --n-cores 8 --wait
Simulation started: scr-4f1c09a2b7d3e815
scr-4f1c09a2b7d3e815: 150/600 s (25.00%)
scr-4f1c09a2b7d3e815: 600/600 s (100.00%)

# download the results to ./fire
$ smoke-cloud download --chid fire
@smoke-cloud/sdk
TypeScript / JavaScript

SDK

Drop the SDK into a Node or browser project and call simulations like any other typed library. .

// npm i @smoke-cloud/sdk
import { ApiClient, KeyAuthProvider } from "@smoke-cloud/sdk";

const api = new ApiClient(
new KeyAuthProvider(keyId, secret),
);

const run = await api.newRun(
{ project: "12345", chid: "fire",
fds_version: "6.10.1", instance_type: 8 },
await readFile("fire.fds"),
);

console.log(run.run_id);
REST API

HTTP API

A clean REST surface for everything the CLI and SDK do — hit the endpoints directly with curl, or use the OpenAPI spec to generate a client in any language.

# run an input file on 8 cores
POST /v5/orgs/{account_id}/runs
?chid=fire&project=12345&fds_version=6.10.1
&instance_type=Cores8
Authorization: Bearer …
Content-Type: application/octet-stream

<the contents of fire.fds>

← 200 OK · run_id scr-4f1c09a2b7d3e815
SDKsTypeScript / JavaScript
AuthOAuth2 · scoped API keys · OIDC SSO
SpecOpenAPI 3.1