Question 1

How does the Stripe connection work — what scopes, and when does data refresh?

Accepted Answer

Dawnset connects to Stripe through  read-only OAuth . We request the bare-minimum scope set we need to estimate MRR impact, surface failed charges, and watch dunning — no write access, no ability to issue refunds, edit products, or move money. Refresh runs once per source per brief — typically a few minutes before   06:14  local. We never cache customer-identifying details, and you can revoke the connection from your Stripe dashboard at any time. Pricing lives on the   pricing page .

Question 2

How does the GA4 connection work, and what events are pulled?

Accepted Answer

GA4 is connected with  read-only property access  via Google OAuth. We pull aggregated acquisition, activation, and funnel drop-off events — never user-level rows, never raw hit streams. The brief compares today’s funnel against your trailing 14-day baseline, so a 14% onboarding drop at step 3 is named against your own history, not against a stock benchmark. Want to see what that looks like in practice? Open the   sample morning brief .

Question 3

Which support inboxes / providers do you support, and what gets read?

Accepted Answer

We support Gmail, Google Workspace, Outlook, and any IMAP mailbox — and Gmail/Workspace through a read-only OAuth scope that cannot send on your behalf. We read subject lines, sender domain, and message bodies to detect refund pressure, escalations, and recurring themes. We do not read mailing-list traffic, marketing threads, or anything outside the inbox you explicitly point us at. Connection tokens are encrypted at rest and revocable in one click from your   settings   page.

Question 4

Which infra dashboards do you connect to, and how?

Accepted Answer

Today: Grafana, Datadog, Better Stack, and a generic Prometheus endpoint. Connection is made two ways, depending on the source: a   read-only API token  you mint yourself (so it can be scoped exactly how you want), or a lightweight   read-only agent  that runs on your infra and pushes a small set of pre-aggregated metrics — error rate, p50/p95/p99 latency, saturation — without opening inbound connectivity. We never ask for write tokens, never request admin scope, and never probe services ourselves. Miss a connector? Reply to your next brief or write to   dawnset@polsia.app   and we will prioritize it.

Question 5

What are the plans, how is proration handled, and what counts as a “seat”?

Accepted Answer

Two tiers today:  Solo  and   Studio . Solo is one founder, one workspace, all four connectors. Studio adds multi-source aggregation across a small team. Full numbers, capping, and feature matrix live on the   pricing   page. A seat is a human email that receives the morning brief or logs into the dashboard — not a connector, not a connected source, not an API consumer. We bill monthly in USD, prorate on upgrade and downgrade to the day, and never annual-lock you in.

Question 6

What happens to my data and my connections if I cancel or downgrade?

Accepted Answer

Because every connection is read-only OAuth or a scoped read-only token, your Stripe, GA4, support inbox, and infra dashboards are exactly as you left them the moment we stop polling them. On the Dawnset side, you can export your briefs, alert history, and tuning state as CSV or JSON at any time from your   dashboard . Request full data deletion and we wipe it within seven days. No proprietary data shape, no migration tax, no lock-in — see the longer answer on the   FAQ .

Question 7

How do I tune sensitivity so I don’t get paged for noise — per source and global?

Accepted Answer

Sensitivity is a single per-source slider (Quiet → Loud, with   Balanced  as the default) plus a global Quiet-hours schedule. Sliders operate on the upstream signal — a stricter Stripe slider ignores sub-$200 dunning misses; a stricter infra slider requires a bigger latency delta before it surfaces. You can also mute any single ranked item by replying “mute” to a brief, which Dawnset remembers and folds into the next day’s ranking. Thresholds and quiet hours live in   Settings . Read about the broader ranking philosophy on the   ranking system   section.

Question 8

What thresholds do you use today, and how do I change them?

Accepted Answer

Defaults are tuned for  seed-stage baselines , not enterprise: roughly a 10–15% deviation from your trailing 14-day median per source, with revenue-weighted overrides for Stripe (any failed-charge cluster > $200 MRR surfaces, regardless of percentage) and latency-weighted overrides for infra (≥ 1.5× p95 for ≥ 10 minutes). Anything that fires above those defaults is shown in the brief with the exact threshold that caught it, so you can dial it. Two ways to tune: (1) the per-source slider in   Settings , or (2) reply to a brief item with the threshold you’d like. We log the change against your account and use it the next morning.

Eight answers before you wire it in.

Integrations

Billing

Tuning alert sensitivity & thresholds

Try the sample brief, or write to us.