Brainy
Design, modify, troubleshoot and validate your cloud architecture using natural language.
This feature is in an alpha stage and open only to select users and customers through a waiting list.
Overview
Brainy is an AI-powered assistant built directly into your Brainboard architecture workspace. It allows you to design, modify, troubleshoot, and validate cloud infrastructure using natural language.
Instead of manually configuring resources, you can simply describe what you want, and Brainy will update your architecture diagram and generate the corresponding Terraform configuration in real time.
Powered by advanced AI models (Anthropic Claude Sonnet 4.6 and Opus 4.6), Brainy acts as an intelligent co-pilot for infrastructure design, helping you move faster while maintaining full control and visibility.
It can take a while for Brainy to analyze the existing architecture, think, and generate results for the requested design configuration. However, you can view continuous progress in both the chat area and the design canvas.
Key Capabilities
1. Real-Time Architecture Editing
Brainy can directly modify your diagram based on your instructions:
Add new resources (AWS, Azure, GCP, etc.).
Update existing components.
Remove unnecessary resources.
Create and manage supporting files (e.g., scripts, IAM policies).
2. Terraform Configuration Management
Brainy automatically handles Terraform configurations, including:
Resource attributes and blocks.
Variables, locals, and outputs.
Structured and clean Terraform code generation.
3. Module Integration
You can use modules directly from your organization’s module catalog.
Reuse predefined infrastructure components.
Maintain consistency across projects.
Speed up complex deployments.
Currently, Brainy can access only the modules that are already imported into Brainboard.
4. Built-in validation
Brainy helps ensure your infrastructure is correct and deployable:
Run
terraform planRun
terraform validateDetect issues and fix them, for example:
Circular dependencies
Resources misconfiguration
Hierarchy problems
5. Pipeline Visibility
Brainy allows you to:
View pipeline execution history.
Access job logs.
Understand what changes are being proposed.
6. Documentation Generation
When requested, Brainy can:
Generate architecture README files.
Update documentation as your system evolves.
7. Context Reference using @
In your prompt/command, you can mention a specific node or a resource that's used in your architecture design by using the @ sign, and you will be presented with the list of the currently opened architecture resources/nodes. Once you start typing the resource/node name, the list will be filtered accordingly for you to choose your desired node.
The node names will appear in the list as the original resource names, and not at the node title.
8. Insights into Brainy's processing
By expanding the Thought option, you can also view Brainy's process or logic, based on which it gave you the results or responded to you.
9. Exporting Chat as a Markdown file
You can export the transcript of your chat with Brainy as a markdown (MD) file. To do so, simply click on the vertical ellipsis (the three dots) in the top right corner of the chat, and select the Export transcript option from the dropdown menu.
10. Redo/Undo changes
Whenever you make changes to your architecture design using Brainy, a clickable Undo option becomes available just below the last Brainy's message from Brainy confirming the change. Clicking on it will revert all the changes that were made up to that point in your current session of chat.
Brainy holds the history of changes that were made within the last 24 hours.
Any changes made before the last 24 hours are discarded, except the last 10 changes that are older than 24 hours. In short, only the last 10 changes older than 24 hours are reversible.
You can also redo/undo changes using a prompt in the chat. However, the 24-hour rule for undo remains the same.
Accessing Chat History
To access the Brainy chat history for any architecture design, simply launch the Brainy chat window, and there you will see the list of past chat sessions. You can also view the time it was created. For example, 30 minutes ago, 10 days ago, etc.
Click on any session that you want to view or continue with.
Limitations
To ensure safety and control, Brainy has the following limitations:
It cannot run
terraform applyordestroy.It cannot access the internet.
It cannot modify
providers.tf(system-managed).It is limited to one architecture per conversation.
It has no access to Terraform state files.
It does not support user/org/permission management.
It is focused only on cloud/infrastructure topics.
It does not have a BYOK / self-hosted option yet (Brainboard team is collecting feedback).
It supports only image files as attachments. Exported chat transcripts (md files), and PDF files are not supported as uploads/attachments as of now.
Data & Privacy
Brainy is built with privacy and security in mind.
What is Shared with the AI
Architecture content (resources, TF configs, variables, outputs).
Conversation messages.
Architecture metadata (name, description, provider versions).
What is NOT Shared
Terraform state files.
Credentials or secrets.
Other architectures.
Personal data.
Zero Data Retention (ZDR)
Brainy uses Anthropic’s Zero Data Retention policy:
No data is stored.
No data is used for training.
Alpha Access
Brainy is currently available in alpha, and the steps for alpha access are:
Join the waiting list.
The feature is enabled via a flag.
Receive documentation and walkthrough.
Provide feedback to maintain access.
Access is revoked if there is no usage for two weeks or no feedback is provided after follow-up.
The feature is free during the alpha preview.
Reporting Issues
When reporting an issue or providing feedback, it is recommended to copy the conversation ID of the chat you used to test Brainy.
To copy the conversation ID, click on the vertical ellipsis (the three dots) in the top right corner of the chat, and select the Copy conversation ID option from the dropdown menu.
Frequently Asked Questions
Is Brainy free? Yes, during the alpha phase.
Can it deploy infrastructure? No. It only supports
planandvalidate.Does it access the internet? No, but it has built-in knowledge of public modules.
Can it make mistakes? Yes. You can:
Ask it to fix issues.
Revert using conversation checkpoints.
Which cloud providers are supported? All providers supported by Brainboard (AWS, Azure, GCP, etc.)
Last updated