GPU-native control plane

Run GPU compute
at full speed

CudaFlow registers your GPU nodes, dispatches simulation jobs, and meters usage — all from a single control plane built for deep-tech compute.

30×

faster than CPU offload solvers

<30s

offline detection latency

0

queue services required

Everything your GPU fleet needs

One API, five agent endpoints, zero external queue services.

Node registration

Agents register over a single HTTPS call with an enrollment token. GPU specs, driver version and compute capability detected automatically.

🎯

Job dispatch

SELECT FOR UPDATE SKIP LOCKED in Postgres — no SQS, no Redis. Jobs are claimed by the fastest available node.

📡

Offline detection

Node liveness is checked at read time against last_seen_at. No background sweepers, no false positives.

📊

Usage metering

GPU-seconds are recorded per job with idempotency keys. Duplicate agent reports are silently acknowledged.

🔐

Auth & orgs

Multi-tenant from day one. Argon2id passwords, session tokens, org membership roles, and enrollment token revocation.

🤖

AI-powered

Claude Sonnet and Opus 5 via Amazon Bedrock — built into the control plane for job analysis and intelligent automation.

Up in minutes

01

Create an org

Sign up, create your organisation, and generate an enrollment token.

02

Register a node

Run the agent on any CUDA-capable machine. It registers itself and starts polling for jobs.

03

Submit a job

POST a job via the API or dashboard. The next available node claims and runs it.

04

Meter & bill

GPU-seconds are recorded automatically. Export usage for billing or analysis.

Ready to run your fleet?

Free to start. No credit card required.

Create your account