Your business teams build their apps with AI,say yes without losing control.
With while, their agent deploys them behind the company sign-in, and you decide who gets in.
Malik’s Claude
- Malik: I’m on the planning team. Deploy my app with while, for the team.
- Malik’s Claude: I’m deploying it, with its database and the morning import.
- Authorize an agentwhile · A new projectAuthorized
- Malik’s Claude: It’s live.Visit scheduleplanning.acme.com
- Malik: Open it to the planning team.
- Malik’s Claude: Requested. The request is waiting for IT’s approval.Open the app to the planning grouprequested by Claude
- Malik: Connect the import to Pennylane.
- Malik’s Claude: I declared Pennylane, read only. It’s waiting for IT’s approval, and I will never see the key.PennylaneRead only · declared by Claude
- while: Known vulnerability in lodashCVE-2021-23337, high
- Malik’s Claude: I’m upgrading lodash to the fixed version, then deploying again.Health / Securitylodash 4.17.21
- Malik’s Claude: It’s fixed and live.
- Malik: What did you do this week?
- Malik’s Claude: It’s all in while’s activity log.
planning.acme.com
Visit schedule
IFMBLDInès, Malik and Léa opened it this morning
A business team’s agent deploys its app
You choose who opens it
The key is entered in while, not with the agent
The agent fixes known vulnerabilities
You see everything the agents do
hello@while.sh
Click your agent to copy its instructions.
Hosted in Franceby Nextmap
What you would have to assemble is already there
Each app opens to whoever you choose, and no one else
With the company account, by group or by domain.
Business teams work on their own, and nothing important happens without you
You choose who approves what, two-person approval included.
You know who did what, and you can prove it
The activity log is signed and verifiable. With the Enterprise plan, it streams to your SIEM.
A business app doesn’t put the others at risk
Each app runs isolated from the others, and the business team’s agent fixes the known vulnerabilities in its dependencies, checked at every deployment, then every day.
The company’s keys stay secret, even from the agents
Apps and agents use your tools without ever seeing the key.
Your data stays in France, with a French company
Your apps and their backups are hosted in France, by Nextmap.
What your security team will ask
- Identity
- Apps open with the company sign-in, over OpenID Connect, with Microsoft Entra, Google Workspace, Okta, Keycloak or any other provider.
- Approvals
- You set rules by role, by group and by app, with two-person approval if you like. An agent never approves.
- Activity log
- It is append-only, chained and signed. It can be exported and re-verified with a small script.
- Isolation
- Each app runs in its own sandbox, with no access to the other apps. Its outbound traffic to the internet is closed by default.
- Known vulnerabilities
- Each app’s dependencies are checked against known vulnerabilities at every deployment, then every day, and the business team’s agent updates them.
- Backups
- Every day, they go to storage outside the cluster where the apps run, dedicated to your organization and encrypted with AES-256. They are kept for 14 days.
- Cutting access
- You disconnect an agent, revoke a key or a token, close a person’s sessions or remove an access, at any time.
- Leaving
- The code of each version is archived. Data can be exported as CSV or as files.
Questions
Getting started
A CFO wants to share their tool: where do we start?
They ask their agent to deploy the tool for their team, and you approve the access.
Do we need a team to run while?
No, there is nothing to install or run on your side.
Governance
How do we oversee the agents of business teams and contractors?
Each agent acts with its person’s permissions, and your rules say what waits for an approval.
Can a business team open an app to the internet without IT?
No, by default, making an app public waits for an admin’s approval.
Who maintains these apps?
The business team’s agent maintains them, within the rules you have set.
What if a business team’s agent breaks its app?
It rolls back to the previous version on its own.
Data
Can the apps use our tools, like the ERP or the CRM?
Yes, each tool connects as a data source, which you approve.
Can one app read another app’s data?
Only after your approval, and read only.
Can an agent delete the database of an app in production?
No: on while, replacing or destroying a production database always waits for a person’s approval.
Can we bring over apps that are on AWS or Heroku?
Yes, the agent takes over the code, copies the database from a backup and carries over the environment variables, as in the guide Leaving Heroku.
Start with one business team’s app
The team’s address
hello@while.sh
Talk to the teamA business team wants to try? They paste these instructions into their agent:
Read while.sh/agents, then deploy my app with while. Copy the instructions