Features

Colleagues can work on shared projects using Claude Code or Codex. Administrators decide who can build an app and who may publish it.

Projects and workflows

Project chat: Write what you need and attach files to the conversation.
Previews and releases: Try an app privately before publishing it. You can roll back a published version.
Workflows: Run work on a schedule or after an event. A run can wait for someone’s approval.

Claude Code and Codex: Use your own supported Claude or ChatGPT subscription, or your own OpenAI or Anthropic API keys.
Connection policies: Set whether project work uses personal accounts, company accounts or either. Automatic assessments use an organisation connection; workflows use an organisation or project connection.
External tools and secrets: Approve MCP tools and GitHub repositories. Choose where encrypted credentials are available: agents, previews or production.

Shared projects: A colleague with access can continue a project using its existing conversation and files.
Roles and permissions: Give a person or team permission to build, publish or roll back. These permissions are separate.
Application access: Let colleagues or the public use an app without giving them access to its source.

Organisation and app SSO: Connect Entra ID, Okta, Auth0 or another OpenID Connect provider. Approved apps can use that sign-in too.
Account security: Use authenticator codes and manage active sessions. Owners have account recovery controls.

AI insights: Check a tool’s usage and running costs. Time estimates include sources and assumptions.
Audit and approval: Record who made a change. Publication can require approval by someone else.
Data checks: Scan supported transfers for secrets. Optional checks for personal data need configuration and do not cover every native provider tool.

Your server: Install on a dedicated Ubuntu server or a supported Kubernetes environment.
Signed updates: Check for signed releases in administration settings and choose when to apply them.
Backups and recovery: Make encrypted backups and test a restore. Keep a separate recovery plan for a lost server.

Choose hosted Software House or install it on infrastructure you manage.

Download a signed release to run on infrastructure you manage. You’ll need a dedicated server or a supported Kubernetes environment.

Choose your hosting plan and where it runs, then review the monthly cost.