# Environments and bots as code

> Keep test and live bots apart with environments, export bots as files, and deploy their settings from your repository with deploy keys.

Environments and deploy keys are in **Settings**, under **Environments**.

![Settings, Environments: Production and Development, and the deploy keys](https://aura.oneinfer.ai/docs/images/settings-environments.webp "Environments, demo data")

## Environments

Every workspace has **Production** and **Development**; you can add up to five in all, such as Staging. Each bot is in one. The **Bots** list on the Overview has a tab for each.

A common setup: a test bot in Development and your live bot in Production, with the same **Key** (set on each bot's **Go live** step, under **Config file**). One config file then deploys both.

A bot in Development costs what one in Production does, $1 a month while it's on. Pause test bots when you aren't testing.

### Lock an environment

**Only deploys change it** locks an environment: its bots' settings change only through your deploy pipeline. You can still pause and turn on its bots, re-sync their knowledge, and set their Slack and Teams webhooks.

## Export and import bots

- On a bot's **Go live** step, under **Config file**, **Export** downloads the bot (its settings, icon and knowledge) as a `.zip`.
- On the Overview, **Export** under the environment tabs downloads every bot in that environment.
- **Import** reads a `.zip` (or a `.json` of settings) and shows what it would change in this environment, bot by bot, before you **Apply** it. New bots are created, at $1 a month each while they're on.

Import is also how you copy bots from one workspace to another.

## Deploy from your pipeline

A deploy key lets your pipeline apply a config file to one environment. A key can read and change bot settings there, and nothing else: no chats, visitors or figures.

1. In **Deploy keys**, give the key a **Name** such as "GitHub Actions", choose its **Environment** and when it **Expires**, then **Create key**.
2. Copy the key now: you won't see it again. Save it as a secret in your pipeline, such as `AURA_DEPLOY_KEY`.
3. In your pipeline, export the environment's bots, plan a change, and apply it:

```bash
curl -sS --fail-with-body https://api.oneinfer.ai/v1/aura/config \
  -H "Authorization: Bearer $AURA_DEPLOY_KEY" > aura.json

curl -sS --fail-with-body -X POST https://api.oneinfer.ai/v1/aura/config/plan \
  -H "Authorization: Bearer $AURA_DEPLOY_KEY" \
  -H "Content-Type: application/json" --data @aura.json

curl -sS --fail-with-body -X POST "https://api.oneinfer.ai/v1/aura/config/apply?commit=$GITHUB_SHA" \
  -H "Authorization: Bearer $AURA_DEPLOY_KEY" \
  -H "Content-Type: application/json" --data @aura.json
```

The Environments page shows these commands with your own address when you create a key.

After a deploy, each bot's editor says when it was deployed, with which key and from which commit. A bot changed in Aura since its last deploy is marked **Changed since deploy**, and the next deploy stops at it until the change is in your file.

## History

Every bot keeps a **History** of its settings, on the **Go live** step: each change from Aura or a deploy. **Restore** sets the bot back to how it was after a change. The restore is itself a new change, so it can be undone too.

## Who can do this

Owners and admins manage environments and deploy keys. Developers can export, import and restore bots. Webhook addresses and saved replies aren't part of config files.