TL;DR: Cloudcraft, Hava and Cloudeval AI all draw cloud diagrams, but they answer different questions. Cloudcraft designs diagrams and, inside Datadog, shows live infrastructure with ops overlays. Hava auto-documents connected AWS, GCP and Azure accounts with version history and change alerts. Cloudeval AI reviews an Azure ARM or compiled Bicep change in the pull request, before deployment.
By Prateek Singh, founder of Ganak AI Labs. Updated 6 October 2026.
"Which cloud diagram tool should we use?" usually gets answered with a feature checklist. That's the wrong frame. Tools that all draw a nice picture often solve different problems.
A better split is the three questions teams ask, at three different times:
- "What does it look like?" You're designing, presenting or onboarding someone.
- "What's running, and what changed?" You're documenting, auditing or debugging the live estate.
- "Is this change safe to ship?" A pull request is open and nothing is deployed yet.
That third job is pre-deployment review: checking a change's design, cost and security before it ships. Cloudcraft, Hava and Cloudeval each lean toward a different question.
Disclosure: I'm the founder of Cloudeval. The Cloudcraft and Hava sections stick to what their own sites say, with links. Prices change, so check before you buy.
What is Cloudcraft best for? Design and live diagrams in Datadog
Cloudcraft is now part of Datadog and comes in two forms.
Cloudcraft in Datadog is a live, read-only view of AWS, Azure, GCP, OCI and vSphere resources (Datadog docs). You can switch overlays for infrastructure, observability, security, cloud cost and APM (AWS only, preview).
The Cloudcraft homepage pitches alerts, logs, security findings and costs on a diagram that updates as your infrastructure changes. Some overlays need other Datadog products, such as Cloud Security for the security overlay. It's free with any Datadog subscription (pricing).
Cloudcraft (Standalone) is a diagramming tool for AWS and Azure, with 3D views of your environments (docs). Two features stand out for design work:
- Live and snapshot diagrams. A live diagram updates from your account. A snapshot is static and fully editable, and you can add components from the Design tab (docs).
- Blueprints, which hold "your diagram, a budget, and all the documentation you attach to individual components" (docs).
Connecting an Azure account to standalone Cloudcraft needs a Pro subscription (docs).
Where it shines: whiteboard-quality design diagrams, budgeting a proposed architecture and, on Datadog, ops context on the live picture during incidents.
What is Hava best for? Auto-generated diagrams and drift alerts
Hava connects to your AWS, GCP and Azure accounts and generates interactive diagrams "grouped by VPCs or resource zones." Its homepage lists:
- A security view of security groups, route tables and connections, to help catch public access to internal services and overly permissive rules.
- Cost forecasts by VPC, region, instance, tag or project. Hava says plainly that it "isn't a replacement for billing platforms."
- Versioning. Hava polls your cloud config, archives old diagram versions and lets you compare them over time.
- Alerts when infrastructure changes.
- Exports to Visio, PNG, CSV or JSON, embeddable diagrams and a self-hosted option.
For automation, Hava publishes a GitHub Action, teamhava/hava-sync-action, to "synchronize and export a diagram after you have deployed changes to your environment". It writes a PNG to your repo.
Every Hava plan includes unlimited diagrams, daily sync and 3, 6 or 12 months of history. The two-week trial includes Teams features, with exports off.
Where it shines: always-current documentation of what's deployed across several clouds, with history and drift alerts.
What is Cloudeval AI best for? Reviewing a change before deploy
Cloudeval can start from the template in your pull request instead of a connected account, before anything is deployed.
Importing a GitHub-hosted ARM template into a project.
- Inputs: ARM JSON, Bicep compiled to ARM JSON, and AWS CloudFormation JSON or YAML as a static beta. For what's already running, live Azure comes in through Cloud sync with a least-privilege, resource-group-scoped custom role (support matrix). There's no live AWS or GCP connection.
-
Output: architecture and dependency views, cost estimates, Well-Architected and validation results, and findings tied to the affected resource. Findings draw on 1,650+ checks. Not every project runs every check, and unsupported areas show as
Not assessedinstead of passing silently. - In the PR: a GitHub Action posts one result comment. It can fail the job on thresholds you set for Well-Architected score, high-risk findings, validation and monthly cost.
- Handoff: the CLI exports diagrams as PNG, JPEG or SVG, which also works for CI or agents.
Posture, critical issues, cost and priorities on one screen.
It isn't a freeform design canvas, an observability overlay or a multi-cloud live inventory. PR diagram diffs and snapshot drift comparison are planned, not shipped yet (interfaces and workflows).
Where it shines: the gap between "the template is valid" and "the template is deployed", when a reviewer wants the shape, cost and risk of a change.
Architecture view with issue badges, dependency labels and cost status.
The difference is the link from a finding to what it affects. This 48-second clip shows it:
How do Cloudcraft, Hava and Cloudeval AI compare?
This maps what each tool is built for, from each vendor's own description. It's not a scorecard.
| Cloudcraft | Hava | Cloudeval | |
|---|---|---|---|
| Primary question | What does it look like? What's happening now (in Datadog)? | What's deployed, and what changed? | Is this change safe to ship? |
| Starts from | Connected accounts (live) or a hand-built design (snapshot) | Connected cloud accounts | A template in a PR, or a live Azure sync |
| Providers (per vendor) | In Datadog: AWS, Azure, GCP, OCI, vSphere. Standalone: AWS, Azure | AWS, GCP, Azure | Azure, plus AWS CloudFormation (beta) |
| Design and editing | Snapshot diagrams, Design tab, blueprints with budget | Not the focus of its main pages | Markup on generated diagrams, not a freeform canvas |
| Change over time | Diagram version history | Version history, diagram comparison, change alerts | Per-commit PR reviews. Drift and graph diff are planned |
| Cost | Blueprint budgets, Cloud Cost overlay in Datadog | Cost forecasts | Monthly cost estimate and an optional budget gate in the PR |
| Security signal | Security overlay (needs Datadog Cloud Security) | Security view of groups, routes and exposure | Findings with evidence from 1,650+ checks |
| CI/CD | Not covered here | GitHub Action and CLI to sync and export after deploy | GitHub Action that comments and gates before merge |
| Pricing (per vendor page) | Free with any Datadog subscription | From $59/month | Free 7-day trial (150 credits), Basic $19/month, Pro $49/month |
The treemap shows where findings cluster by resource type.
Can you use Cloudcraft or Hava with Cloudeval AI?
Yes, and that's often the most practical setup. This pipeline reviews on the pull request, then refreshes documentation diagrams after deploy.
# .github/workflows/infra.yml (sketch: wire in your own deploy job)
on:
pull_request:
push:
branches: [main]
permissions:
contents: read
pull-requests: write
issues: write
jobs:
pre-deploy-review: # Question 3: is this change safe to ship?
if: github.event_name == 'pull_request'
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: ganakailabs/cloudeval-action@v1
with:
access_key: ${{ secrets.CLOUDEVAL_ACCESS_KEY }}
project_id: ${{ secrets.CLOUDEVAL_PROJECT_ID }}
mode: review
post_pr_comment: true
upload_artifacts: true
# deploy: ... # your existing deployment job
refresh-docs-diagram: # Question 2: what's deployed now?
if: github.event_name == 'push'
# needs: deploy
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: teamhava/hava-sync-action@v1
with:
source_id: ${{ secrets.HAVA_SOURCE_ID }}
hava_token: ${{ secrets.HAVA_TOKEN }}
environment_id: ${{ secrets.HAVA_ENVIRONMENT_ID }}
view_type: infrastructure
image_path: docs/architecture.png
# commit docs/architecture.png in a follow-up step if you want it in the repo
The Cloudeval job needs the repo linked through the Cloudeval GitHub App. The Hava job needs a connected Hava source. On Datadog, Cloudcraft's live diagram replaces the second job, outside CI entirely.
Which cloud diagram tool should you choose?
Ask these in order:
- Designing something new and presenting it? Cloudcraft's snapshot diagrams and blueprints.
- Already paying for Datadog? Cloudcraft in Datadog is included.
- Need always-current docs and change history across AWS, GCP and Azure? That's Hava's core pitch.
- Changes reach production through pull requests, and nobody checks design, cost and exposure there? That's the gap Cloudeval targets.
Some teams answer "yes" to two. That's fine, because the tools mostly don't compete for the same moment.
Frequently asked questions
What is a good Cloudcraft alternative for Azure?
It depends on the job. Hava is built for always-current documentation and change history of deployed Azure, AWS and GCP accounts. Cloudeval AI reviews an Azure ARM or compiled Bicep change in a pull request before it ships. See the longer Cloudcraft alternative for Azure pre-deployment review.
Does Cloudcraft support Azure?
Yes. Per Datadog's docs, Cloudcraft (Standalone) covers AWS and Azure, and connecting an Azure account needs a Pro subscription. Cloudcraft in Datadog covers AWS, Azure, GCP, OCI and vSphere.
Is Cloudcraft free?
Cloudcraft's pricing page says it's free with any Datadog subscription, with unlimited users and diagrams. Standalone Cloudcraft has Free, Pro and Enterprise plans, and connecting Azure needs Pro.
How much does Hava cost?
Hava's pricing page lists Professional at $59/month (3 cloud accounts, 1 user), Teams at $249/month (10 accounts, unlimited users) and custom Enterprise pricing. Check the page before you buy, because pricing changes.
Can you use a diagram tool and a pre-deployment review tool together?
Yes. They run at different moments: review the change on the pull request, then refresh documentation diagrams after deploy. The workflow sketch above runs both from one GitHub Actions file.
Takeaway
"Cloud diagram tool" covers at least three jobs: designing, documenting what's live and reviewing a change before it ships. Pick the tool by the question you need answered, and when.
A diagram of production won't warn you a PR is about to double the bill. A PR gate won't help you whiteboard next quarter.
If pre-deployment review is the missing piece, read the criterion-by-criterion comparison of Cloudcraft and Cloudeval for Azure pre-deployment review.
Related on cloudeval.ai
- Compare Cloudeval with other cloud review and diagram tools
- Hava.io alternative for reviewing Azure changes before deployment
- Lucid Cloud Accelerator alternative for Azure pre-deployment review
- How to generate an Azure architecture diagram automatically
Try it
Want to see a template as a reviewed diagram before it ships? Open the Cloudeval app and follow the diagrams and workspace views guide.
About the author
Prateek Singh is the founder of Ganak AI Labs and builds Cloudeval AI.
cloudeval.ai | Docs | YouTube | GitHub | LinkedIn
Top comments (0)