Massdriver

Massdriver is an internal developer platform that turns a platform team's Terraform, OpenTofu and Helm into a policy-governed catalog, SaaS or self-hosted.

Massdriver gives developers and AI agents a governed path to the cloud infrastructure a platform team has already defined. Platform engineers publish the Terraform, OpenTofu and Helm they write to a catalog, and everyone else provisions from it with policy enforced. The product works as an internal developer platform and a platform orchestrator, running existing infrastructure code rather than replacing it with a new language. Massdriver, Inc., a remote-first company, builds it and sells it as one platform. Infrastructure code is packaged as bundles, versioned with semantic versioning and published to the organisation's catalog. Provisioners run OpenTofu, Terraform, Helm and Bicep, and custom provisioners run as Docker images. Bundle instances link through connections, so one instance's resource fills a slot another bundle declared. Each change travels as a proposal with its plan diff and results attached, and compliance policy checks run on the plan before apply. Attribute-based access control on environment, team and blast radius decides who may plan, propose or deploy. State sits in a built-in HTTP backend with locking, existing resources come under management without a rebuild, and Cost Insights reports spend by resource, environment and project. Platform engineering teams that already maintain Terraform, OpenTofu or Helm get the most from it, because their existing modules become the catalog. Organisations letting AI agents touch infrastructure use it so that agents provision from the same catalog under the same policies as people. Citizen developers on product teams get approved building blocks instead of raw cloud consoles. Regulated companies can install the whole platform on their own Kubernetes cluster with Helm, including in air-gapped networks, with SSO and SCIM for enterprise sign-in. A typical job starts when a developer picks a bundle from the catalog, connects it to an existing instance and submits the change as a proposal. The platform team reads the plan diff, the policy results arrive with it, and an approver releases the deployment, from the platform or with mass deployment approve in the CLI. Release pipelines in GitHub Actions or Azure DevOps drive the same workflow, and AI coding agents such as Claude Code work through the catalog too. Massdriver fits teams that want every infrastructure change, whether a person or an agent made it, to pass one review before it reaches a live environment.