Manage your integrations
Integrations are the connections CloudQuell reads your billing and usage data from. You manage them in one place — add new connections, keep existing ones healthy, and remove any you no longer need.
What you can connect
Section titled “What you can connect”- AWS — live and connectable today.
- Snowflake — connectable today via read-only key-pair auth. See Connect Snowflake.
- Anthropic — connectable today via a Claude Console admin key. See Connect Anthropic.
- OpenAI — connectable today via a read-only Admin API key. See Connect OpenAI.
- Microsoft Azure — built and connectable on request (limited availability). See Connect Microsoft Azure.
- Google Cloud — built and connectable on request (limited availability), keyless via Workload Identity Federation. See Connect Google Cloud.
CloudQuell reads AWS spend through a cross-account, read-only IAM role and a Daily, Parquet CUR 2.0 you deliver to your own S3 bucket. You paste the role ARN; CloudQuell assumes it with an ExternalId, and no AWS credentials are stored. Cost Explorer backfills about 12 months of history the moment you connect.
The full first-time setup — trust policy, inline policy, CUR delivery — is documented in Connect your AWS account.
You can connect several AWS accounts. Connecting from a management (payer) account automatically covers its linked accounts.
Snowflake
Section titled “Snowflake”CloudQuell reads Snowflake spend through read-only key-pair (RSA) authentication. A
short setup script — which you run once as ACCOUNTADMIN — creates a read-only
CLOUDQUELL service user, a cost-capped XS warehouse, and owner’s-rights wrapper views
over Snowflake’s ORGANIZATION_USAGE billing views. CloudQuell generates the key
pair and stores the private key in AWS Secrets Manager — never in the database,
and you never handle a private key. Connecting requires an ORGADMIN-enabled
organization account. About 12 months of history backfills on connect; after that
spend syncs daily, with dollars finalizing up to ~72 hours behind to match
Snowflake’s own end-of-month adjustments.
Rename, edit, and manual re-sync work the same as any connection. The full setup — account identifier, setup script, verification, and troubleshooting — is documented in Connect Snowflake.
OpenAI
Section titled “OpenAI”CloudQuell reads OpenAI spend through a read-only Admin API key (sk-admin-…)
that you paste at connect time. The key is validated live against OpenAI, stored in
AWS Secrets Manager — never in the database — and can be rotated any time from
the integration’s Manage dialog. About 12 months of history backfills the
moment you connect; after that, spend syncs daily.
The full setup — key creation, what data appears, troubleshooting — is documented in Connect OpenAI.
Anthropic
Section titled “Anthropic”CloudQuell reads Anthropic spend through a Claude Console admin key
(sk-ant-admin01-…) that you paste at connect time. The key is validated live
against Anthropic, stored in AWS Secrets Manager — never in the database — and
can be rotated any time from the integration’s Manage dialog. Claude API usage
and cost sync daily into the AI Spend view.
The full setup — admin-key creation, what data appears, troubleshooting — is documented in Connect Anthropic.
Manage a connection
Section titled “Manage a connection”For any live connection you can:
- Rename it so it’s easy to identify.
- Edit its settings — CUR path, role ARN, or cost-display options.
- Trigger a manual re-sync to pull fresh data on demand.
Remove a connection
Section titled “Remove a connection”Removing a connection is a permanent hard delete. It removes that connection and everything ingested for it — cloud-side ingest resources, stored files, and query partitions. This can’t be undone, so it requires confirmation and is admin-only. For Snowflake, also run the teardown script in your account to drop CloudQuell’s read-only objects.