> ## Documentation Index
> Fetch the complete documentation index at: https://docs.aient.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent-assisted setup

> Use an AI assistant to prepare Aient setup safely: choose a plan, gather repository details, connect approved integrations, and install telemetry without bypassing human consent.

# Agent-assisted setup

An AI assistant can help you prepare Aient setup, but it should not bypass payment, integration consent, or code review. Use it to gather the right details, follow the docs, and draft the changes you review.

## What an assistant can do

| Task                       | What to ask for                                                                                                                           |
| -------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
| Choose a setup path        | Compare the server, browser, and source map guides for your stack.                                                                        |
| Prepare repository details | Identify the service name, deployment environment, and repository to monitor first.                                                       |
| Install telemetry          | Apply the [server](./quickstart/server) or [browser](./quickstart/browser) quickstart and keep release metadata aligned with deployments. |
| Add source map uploads     | Wire the [source map upload](./source-maps/upload) command into CI after the frontend build.                                              |
| Summarize next actions     | Produce a checklist for payment, GitHub, Slack, telemetry keys, and verification.                                                         |

## Human approvals

Aient keeps high-trust actions behind a human or workspace admin:

* choosing or changing a paid plan
* approving payment in Stripe
* installing the GitHub App
* connecting Slack and selecting channels
* creating or rotating server-side credentials
* reviewing and merging remediation PRs

If an assistant reaches one of those steps, it should hand you the link or checklist and wait for confirmation.

## Information to prepare

| Detail             | Example                                             |
| ------------------ | --------------------------------------------------- |
| Organisation name  | `Acme Engineering`                                  |
| Repository         | `acme/web`                                          |
| First service name | `checkout-api`                                      |
| Environment        | `prod`, `staging`, or `dev`                         |
| Runtime            | Node.js server, browser app, or both                |
| Release metadata   | commit SHA, branch, deployment environment          |
| Slack channel      | the channel where Aient should post problem updates |

Stable names matter. Use the same `serviceName` in telemetry and source map uploads so Aient can connect stack traces to the right release.

## Safe credential handling

* Use publishable environment keys in application telemetry.
* Store source map upload keys only in CI or other server-side secret stores.
* Do not paste private keys, Stripe secrets, GitHub tokens, or Slack tokens into a general-purpose assistant.
* Ask the assistant to make a PR or patch, then review the diff before merging.

## Verify setup

After the assistant makes changes:

1. Deploy a small change with release metadata.
2. Confirm traces and logs arrive in Aient.
3. Trigger or inspect a known error in a non-production environment first when possible.
4. Upload source maps for the same commit SHA.
5. Check that stack traces resolve to readable files and line numbers.

## Related guides

* [Server telemetry quickstart](./quickstart/server)
* [Browser telemetry quickstart](./quickstart/browser)
* [Upload source maps](./source-maps/upload)
* [OpenTelemetry endpoints](./reference/otlp-endpoints)