Migration guide

Switching from a metered helper.

You have a tool that sells AI help by the credit or the hour, and a reason to be reading this. Here is how to move without a leap of faith: run the roster in parallel at the drafting level, compare artifacts, and cancel nothing until the evidence says so.

The method

Five steps, none of them a cutover.

1. Read your last invoice
Before comparing products, compare outcomes: list what last month’s credits or hours actually shipped. Not tasks started — work that left the building. That list is your baseline, and it is usually shorter than the invoice suggests.
2. Map the work to names
Content and social land with Owen, growth research with Marisol, support drafts with Sana, the books with Lena. Whatever your current tool does not touch — planning, procurement, infrastructure, audit — is roster work too; it just was not on the old menu.
3. Run in parallel at T1
Everyone starts as a drafter: real work on your real inputs, parked for your release. Nothing sends, spends, or changes anything until you release it, so running both systems for a month carries no cutover risk — and at the starter profile the parallel month costs $39 plus single-digit usage.
4. Compare artifacts, not promises
At month’s end you hold two piles of work product. Read them side by side against step 1’s baseline. This is the whole comparison — no demo, no sales call, no benchmark of ours.
5. Promote, then cancel
If the parked drafts earn it, promote a rung at a time — each promotion your decision, on the ledger with a reason. Cancel the meter when it has nothing left to do. If the drafts do not earn it, cancel us instead: the fee is refundable in full for 30 days, no questions.

What changes

The three differences you will feel in week one.

Parked, not sent

A metered helper acts and you audit afterward. A drafter parks and you release. The first week feels slower and reads better — you see everything that would have gone out, before it does.

A cap, not a meter

Nothing here runs out mid-month. Inference is billed by your provider at their prices under a hard cap you set — enforced before the spend, not billed after it — and the fee is $39 regardless.

A record, not a history

Every permission the roster holds is on an append-only ledger with the reason it was granted. When you widen something, that is on the record too. Your old tool’s chat history is not this.

The specifics

If your current tool is one of these.

The dated, linked specifics — prices, meters, oversight models, and what each product is genuinely good at — are on the comparison pages. Every figure there carries the month we verified it.

vs. Sintravs. Marblismvs. Lindy

Direct answers

The switching questions.

Do I have to cut over all at once?

No, and you should not. T1 exists so that the roster proves itself on your real work while your current arrangement keeps running. The switch happens promotion by promotion, each one your call.

What if the roster is worse at something my current tool does well?

Then keep the tool for that thing, or keep the person — roles move one at a time here, and we do not need to win them all. The comparison pages say plainly where a competitor is the reasonable buy.

What does the parallel month cost?

$39 plus usage at your provider’s prices — single digits to about $15 at the starter profile, as an estimate. If the month does not earn the switch, the fee is refundable in full for 30 days.

Run the parallel month.

A working session, your keys, and a month of parked drafts against your own baseline. The meter can wait for the evidence.

Request an inviteWhat the total costs