Launched this week
Cloud World Model
Simulate AWS, GCP & DigitalOcean without paying the bill
288 followers
Simulate AWS, GCP & DigitalOcean without paying the bill
288 followers
Simulate AWS, GCP, Azure, OCI & DigitalOcean architectures to predict cost, performance, and resilience without provisioning real resources or paying a cloud bill. Built for learners practicing cloud skills and AI agents training on cloud optimization.
The agent-training angle is interesting. I’ve wasted time spinning sandbox infra just to test small failure cases, so having cost + autoscaling + failure recovery in one sim feels useful. Curious how you keep provider-specific limits and pricing drift up to date?
Cloud World Model
@xiaosong001 We use CI to keep it up to date and fail the build if its not. We also have weekly jobs that runs a pricing benchmark for all five providers and fires an alert to our dashboard if any number crosses a drift threshold.
This is the kind of tool I wish existed when I was doing architecture reviews - being able to say 'here's what we're planning to build' and actually simulate the traffic/cost without provisioning anything is huge. Does it model network latency between regions, or is it mostly a cost estimator at this point?
Cloud World Model
@omri_ben_shoham1 Thanks Omri, that's exactly the use case we built this for. Architecture reviews where you can say "here's what we're planning" and get real behavioral signal before anyone writes an infrastructure ticket or even finishes creating the architecture documentation for that matter.
To answer your question: it's well beyond a cost estimator. The simulation engine models latency curves (including cross-region network latency between provider regions), CPU saturation, error rates, throughput, connection pool pressure, and autoscaling behavior all simultaneously. The hybrid prediction engine combines deterministic rules with ML-based prediction to surface things like scaling delays, provider-specific quirks, and bottleneck forecasts.
Would love to hear what kind of architectures you typically review, always useful feedback for what to prioritize next.
Cloud World Model
@saaduddin_ansari love this feedback, and yes, this is exactly the direction we're building toward! The LocalStack comparison is spot on for how we think about it too.
Right now we have a TypeScript SDK (OpenAPI-generated) that wraps our simulation API, so you can drive simulations programmatically today. The go-sdk drop-in replacement you're describing, where you'd swap your AWS/GCP SDK endpoint to point at Cloud World Model and get realistic simulated behavior back is on our roadmap. It's the most requested integration pattern we've heard.
The core simulation engine already models AWS, GCP, Azure, OCI, and DigitalOcean behaviors (latency curves, autoscaling quirks, pricing, failure modes), so the behavioral fidelity for a drop-in is already there. The missing piece is the API surface compatibility layer so existing SDK calls route through seamlessly.
Would a Go client for our current simulation API be useful in the meantime while we work toward full SDK compatibility? Happy to prioritize that if it unblocks your workflow.
Really cool - had a multi-cloud setup simulating in a couple minutes. One thing though: connecting resources took a few clicks each time (here's what I mean: Cloud World Model | createademo, ~0:10). Is there a faster way to wire them up - a drag-from-handle or keyboard shortcut? Would speed up building a setup a lot. Congrats on the launch!
Cloud World Model
@john_marker3 That's awesome that you was able to create a quick demo. Truthful, I actually love the API more than the UI. I actually considered creating no UI but learners need a UI. The quickest way is to use the predefined scenarios on the scenario tab. Your point thought is valid regarding a quicker way to wire them up. Will need to think that through a little bit more. Thank you for the feedback, for creating the demo and the congratulations comment.
Cloud World Model
@john_marker3 Issued a UI fix for this. Curious on your feedback when you have a moment.
Here's how you validate:
1. Open the Workspace (the "Workspace" tab at the top, which is already showing).
2. Add two resources to the canvas (example - 2 web servers and a load balancer or servers and databases)
Hover over one of the nodes. Four small round handles should appear on its edges -top, right, bottom, and left.
Press down on a handle and drag. A thin "ghost" line should follow your cursor.
Drop it on the other node. When you hover over a valid target, it highlights; releasing creates the connection line between the two nodes.
Confirm the old way still works, the "Connection Mode" button should still create connections the same as before.
Curious, if this mechanism is faster. Thanks, in advance. You're feedback was great!
@mathsociety Just tried it. That's definitely faster and feels much more natural. I think that's a BIG improvement over the previous workflow. One other thing that would be awesome is being able to freely move the objects around on the canvas after placing them (see the video for what I mean). I think that would make building diagrams feel even smoother. Nice update! https://createademo.com/v/cmr2nge290001lb046hoc7efu
Cloud World Model
@john_marker3 The move the objects around canvas issue is now resolved as well. It was an issue on Chrome. It's now fixed on Chrome. Thanks!!
Macaly
simulating the cloud bill away is realy clever 🙌 fidelity is the whole game tho
Cloud World Model
@petrkovacik yes, we believe we have fidelity. https://www.cloudworldmodel.ai/fidelity
@mathsociety I like this because cloud mistakes usually get expensive after you already deployed them. Being able to play with failure scenarios, cost, and scaling before touching real infra feels especially useful for learners and small teams that don’t have a proper staging setup.
Cloud World Model
@alpertayfurr Thank you. I’m glad you recognize this. We see Cloud World Model as something for human learners and non-human learners. It’s expensive to burn real money or to have an incident. I could remember we used to spend weeks on architecture. The only way we was able to validate architectural decisions was to put it in the environment or best practice. Stress testing an architecture is cost especially for small teams, even for large teams.
Exactly, that validation step is usually the expensive part. Simulating the tradeoffs before touching real infra feels like it could save both money and a lot of painful guesswork.