Same engine on every plan. More hands-off as you scale.
What actually changes between tiers.
Between tiers
Pricing
Customer limit
Your role
Autonomous sending
Weekly dividend
Channels
Multi-message arc sequences
Warehouse sync · Snowflake / Databricks
Full audit-log export via API
Priority support · dedicated account management
Seed Free forever
Autonomous · release all
Free
Up to 1,000 customers · no card
The engine scores, drafts, holds until the window opens, and queues every decision with its reason. You release the queue in a click. Full timing engine from day one, no approval gate, no reduced model.
1,000
Release all
Off until you switch it on
✓
Email
—
—
—
—
Growth 30-day trial
Autonomous · exceptions only
$499
per month · up to 10,000 customers
It sends on its own when the read is strong enough, surfaces the borderline accounts for you, and holds the rest until their window opens. You touch the exceptions, not the queue.
10,000
Exceptions only
Off until you switch it on
✓
Email + WhatsApp
✓
—
—
—
Scale 30-day trial
Autonomous · outcomes only
$1,499
per month · up to 50,000 customers
Full autonomy inside the timing engine, with a value threshold above which the engine hands the account to your team instead of sending. Every decision logged and cited. You review outcomes in the ledger, never approvals.
50,000
Outcomes only
Off until you switch it on
✓
Email + WhatsApp + SMS
✓
✓
✓
✓
A first-customer brand on Seed runs the same engine as a brand on Scale. What changes is volume and your role on the exceptions the engine surfaces, never the engine itself.
Every plan reports where your board reads it: LTV per cohort and NRR, not open rates.
Same price either door: start fresh on BHASM, or keep your stack and let Imprint learn how you run retention today.
Numbers we can stand behind.
Evidence
Calibration
Nothing to configure
80 industry profiles ship with their own rhythm and sensitivity tier, calibrated before you connect a single data source.
Capacity
A million scores, 14.4 seconds
On a single core. Volume is not the thing that constrains this.
Windows
Twelve of twelve
Twelve of twelve buying windows derived on their own, 207 to 234 days, controlled dataset.