AI diagram generator

AI AWS architecture diagram generator.

From the Terraform you already have, with no read-only role to provision and nothing sent anywhere.

Download for macOS
v0.1.33 · .dmg · Apple Silicon & Intel
Signed & notarized by Apple · opens without a Gatekeeper warning
sha256 698955a0187bc039f4c74f5d05a9f10fbb27376a45788a0a241d1326b73873c7
Download for Windows instead
$curl -fsSL https://lucidtrain.com/install.sh | sh

Sources it reads

Terraform, CloudFormation templates, CDK synthesised output, SAM templates, and JSON from describe commands if that is what you have. Each of them fully describes the infrastructure, and each of them is a file on disk, which is why none of this requires access to the account it describes.

How Lucid Train draws one

Describe what you want, or point it at a repository. The agent emits a semantic graph of nodes and edges with no coordinates, and the ELK layout engine computes positions and orthogonal routing. That separation is the reason the output cannot come back with boxes overlapping or arrows crossing through cards, which is the characteristic failure when a language model is asked to place things on a canvas itself. It also means editing is editing the graph: change a component and the layout re-solves rather than drifting out of alignment.

Tiers rather than VPC boxes

The default AWS diagram convention nests everything inside VPC and subnet rectangles, which spends most of the picture on network structure. Grouping by tier instead puts the emphasis on what the system does, with network placement as an annotation. If the network genuinely is the subject, that is a different diagram and worth drawing separately.

Queues as edges, not boxes

SQS, SNS and EventBridge are drawn as the connection between producer and consumer rather than as nodes traffic passes through. That keeps the component count down and matches how people describe these systems out loud, where nobody says a request visits the queue.

Local, offline, your models

It runs as a desktop application. Against a local Ollama model the whole thing works offline with no API key and no telemetry, which is the practical difference from every browser-based diagram tool: your architecture never leaves the machine.

AWS services by tier

ServiceTier
CloudFront, Route 53, WAFEdge
API Gateway, ALB, NLBEdge
EC2, ECS, EKS, FargateApplication
Lambda, Step FunctionsApplication
RDS, Aurora, DynamoDBData
S3, EFS, FSxData
ElastiCache, MemoryDBData
SQS, SNS, EventBridge, KinesisDrawn as edges
Cognito, IAM, Secrets ManagerAuthentication boundary
CloudWatch, X-RayInfrastructure

Try it

shell
$ Read the Terraform in this directory and draw the AWS architecture. Group by tier, show queues as edges, and note which resources are multi-AZ.
FAQ

Questions

Does it need AWS credentials?

No, and it never asks for them. Everything comes from files.

Does it work with CDK?

Yes, from the synthesised template. Reading CDK TypeScript directly works too but is less reliable, since the template is what actually gets deployed.

Can it show multi-region architecture?

Yes, with regions as sections. Cross-region edges are worth marking explicitly since they are where latency and data transfer cost accumulate.

How does this compare with the AWS architecture icons in draw.io?

Those give you a palette and you do the drawing. Here the diagram is generated from your infrastructure code and laid out by a solver, so it starts from what you actually deployed rather than from a blank canvas.

Related