Olyv positions you, finds the people worth writing to, drafts the sequence and sends it on a schedule — stopping the moment someone replies.
No sign-up. Load the worked example and click through the whole pipeline.
Each stage is constrained by the one above it. You make the first call; the agent carries it the rest of the way without asking you again.
Four routes. You pick one.
Genuinely different routes — different buyers, different proof, different competitors. Four more on request, up to sixteen. The one you choose constrains everything below it.
You decide this. Nothing else asks for your time.
The segment becomes real people.
An ICP with disqualifiers, not just fits — most wasted spend comes from a list that excluded nobody. Then named prospects with verified addresses.
Every prospect records which provider supplied it.
A sequence, sent on a schedule.
First touch, follow-ups, break-up — all making the same claim. Sent inside working hours in the prospect’s own timezone, at a rate the domain can carry.
The moment someone replies, their sequence stops.
Replies triaged. The rest come to you.
Interested, not now, wrong person, unsubscribe. Clear positives get a meeting booked; anything read with low confidence is escalated with the thread.
This is the edge the agent stops at, and it stops visibly.
What stops it
The question about a tool that sends while you sleep is never whether it can send. It is what stops it doing something expensive at three in the morning. Here is the whole list — every one of them is a plain function in the repository, and every refusal is written onto the prospect it stopped, in words.
Four are permanent: that prospect is finished. Three are about timing: it waits and tries again. The difference is recorded, because a campaign that quietly sends less than you expected is a campaign nobody can debug.
They replied
Stops for goodThe sequence stops immediately, and this is checked before anything else — including whether the campaign is even running. A follow-up to someone who already answered is the fastest way to lose them.
They asked not to be contacted
Stops for goodTerminal, permanently, no exceptions and no re-enrolment in a later campaign.
The address bounced
Stops for goodSending again costs the domain far more than one prospect is worth.
The data provider is unsure of the address
Stops for goodBelow 70/100 confidence it is not attempted. A bounce is more expensive than a missed prospect.
It is outside their working hours
WaitsEight to five in the prospect's own timezone, not the server's. Weekends off unless you say otherwise.
The daily cap is reached
WaitsThirty a day by default. A new domain that opens at volume stops being delivered, and no campaign is worth that.
No sending identity is connected
WaitsNothing is attempted from an unconfigured domain, and the campaign says so rather than failing quietly.
Questions
A third-party contact data provider — the same kind everyone in this category uses. Olyv has no database of its own and does not pretend to; every prospect records which vendor supplied it and their id for it, so a record can be re-resolved or disputed later. The quality of your list is the quality of that provider, and swapping to a better one is a configuration change.
Yes, through a sending provider you connect. That means deliverability is your domain's reputation, and Olyv does not repair a domain that is already burned — no tool does, whatever it claims. The defaults exist to protect a healthy domain rather than rescue a damaged one: thirty a day, working hours only, weekends off.
A unique constraint in the database on the pair of prospect and sequence step. The runner is written to be idempotent, but designed-to-be is not the same as cannot — two overlapping scheduled runs, or a retry after a timeout that actually succeeded, would otherwise be enough. The send row is also written before the provider is called, so a duplicate attempt loses that insert and stops having sent nothing.
Their sequence stops on the spot. It is the first condition checked, ahead of whether the campaign is even running, so a reply arriving during a pause or out of hours still ends the sequence. The reply is then classified — interested, not now, wrong person, unsubscribe — and anything the system cannot read confidently goes to a person with the thread attached.
Most of them are volume machines, and volume is genuinely table stakes now. The difference is that Olyv decides its positioning first and generates every message in every campaign from that one decision, so a campaign can be read backwards — any email, at any time, names the route it came from. That matters the day a prospect gets two emails from you making two different arguments.
Nothing yet, because there is nothing to charge for yet. Olyv is pre-launch: the workspace runs free with no sign-up, and the outbound engine runs in dry run until you connect your own provider keys. When pricing exists it will be on this page rather than behind a call.
No. Your positioning, prospects and replies are yours; nothing is pooled across accounts and nothing is used to train a model. Every table is scoped to your organisation with row-level security in the database, not only in application code.