Security and Privacy at Bunting Labs

Last Updated: February 21, 2025

Our customers' security and privacy is important to us. We've designed multiple deployment options that can be adopted to meet different security and privacy requirements.

In this document, we outline:

  • how customer data is shared with third parties when using our services
  • how choosing different tiers of our services changes data sharing

To contact us about our security and privacy practices, or to request data erasure, please email us at support@buntinglabs.com.

Personal Cloud Professional Cloud Organization Private cloud / on-premise
Third Party Data Providers
OpenAI & Google
AI models
Sees map data Sees map data Local alternative available
Cloudflare
Data storage
Sees map data Sees map data Local alternative available
Stripe
Payment processor
Sees billing info Sees billing info Sees billing info
PostHog
Analytics
Sees account info Sees account info Sees account info
Resend
Transactional email
Sees account info Sees account info Sees account info
Stytch
Authentication
Sees account info Sees account info Sees account info
Data Retention
Chat history One year Deleted after chat session Deleted after chat session
Data erasure Upon request Upon request Upon request

Frequently asked questions

What's the local LLM alternative?

For organizations that are deploying on a private cloud environment, we support using many open source AI models, like Mistral or Llama. Using a local LLM means no third party LLM provider can see your data.

What map data is sent to the server when using Kue AI?

Kue sends chats along with metadata describing your layers, including the first row of the attribute table. If you drag a layer into the chat window, the entirety of the layer data is uploaded to our server. This is required for metadata enrichment or making a web map.

What does an on-premise deployment look like?

We'll provide a Kubernetes Helm chart to deploy our software. You can deploy that on your own bare metal or a cloud provider like AWS or GCP. Hardware requirements are approximately 4 vCPUs, 8GB of RAM, at least 2TB+ of disk space, and a GPU for local LLMs inference.