Three tools, one job: say the hard thing well.

Respond and Read are usage-light by design: 100 drafts and 100 analyses per billing period. Those are soft caps — go over and you get a heads-up, not a locked door. Track holds up to 25 active clients.

The first time you open Respond it asks for two to five emails you have really sent to clients. It reads them for your greetings, your sentence length, how direct you are, how you sign off — and applies that to every draft afterwards. Skip that step and the drafts still work, they just sound like an AI wrote them.

Pick ordinary working emails: a status update, a nudge about an invoice, a reply to a scope change. Anything you wrote for a stranger, copied from a template, or made unusually formal will teach it the wrong voice. Your samples are stored so you can edit them later, are never shared, and are never used for training. You can delete the whole profile from settings in two clicks.

Paste the message that made your stomach drop, add one line of context if it helps, and Respond picks a strategy mode and drafts the reply. If it picked wrong, tap a different mode and the draft rewrites itself.

Every draft is written at one of five levels of firmness. They escalate in order — clarify is the lightest touch, exit is the last one. Respond will suggest the lowest mode that actually addresses the situation, because opening at defend when clarify would have worked is how good clients get lost.

Paste a message and Read gives you three things: a tone score from 0 to 100 measuring how adversarial it is, a list of the specific pressure patterns it found with the phrase that triggered each one, and the mode it thinks you should answer in — with the reasoning.

The point is not the score. The point is the flags: the sentence that quietly moved a deadline, the compliment doing the work of a scope add, the phrase that reads friendly and commits you to a second round of revisions. Once you can see it, hand it straight to Respond and the draft starts from what Read found.

Track never touches your inbox. You log moments as they happen — one tap, under ten seconds — and it keeps a running health score per client. Every client card can be opened to see exactly which rules fired, what each one was worth, and why the number is what it is. There is no AI anywhere in the scoring: it is a fixed set of rules, so the same history always produces the same score.

Everyone starts at 80. Paying fast, paying at all, and long stretches with nothing logged move it up; unpaid invoices, repeated scope adds, off-hours pings, revision churn and tone shifts move it down, weighted by how recent and how severe they are. 70 and above is good, 45–69 is drift, and below 45 is at risk.

Drift is the band that earns the subscription. It is the stage where the relationship is still fixable and you would not have noticed on your own — which is exactly when a clarify or align reply is worth sending.

Each tool bills separately and monthly. Cancel any of them from Billing and it stays live until the end of the period you already paid for — no exit interview, no retention offer. Your voice samples, your pasted messages, and your client log are yours: none of it is shared, sold, or used to train a model.

Something behaving oddly, or a draft that missed badly? Reply to any email from us and it reaches a person.