Why we exist
Before writing a line of Rainmaker, we reverse-engineered the two most-hyped AI SDR tools in the market from inside their own apps. Not as critics. As users, trying to see if either could do what it claimed.
What we found: one charges per seat, so the price grows with your headcount, not your results. The other's demo silently fails for anonymous users, and its ICP engine returned our competitors as our ideal customers. Neither cites a source for any claim in any email. Both run on autopilot with no hard stop: when the model cannot back a claim, the email goes out anyway. The observations are dated and written up in full on Rainmaker vs Gojiberry and Rainmaker vs Explee.
We built Rainmaker to be the opposite of both. Every factual claim must be sourced, every unsourced claim is a hard stop the agent cannot talk its way past, and pricing is partly tied to meetings the prospect accepts and keeps past the scheduled start. Not seats, and not sends.
Meetings you can defend.
Not send volume. Not open rates. Every product decision runs through one question: does this help the user book meetings they can genuinely defend? If it optimizes for a number that isn't that, we cut it.
Quality over quantity
Fewer, better-researched touches that cite real evidence. A bigger list is not a strategy.
Autonomous, with a hard stop
The agent finds, writes and sends inside the rules you set. The moment a claim cannot be sourced, it stops and waits for you - and nothing can release that draft but you.
Verifiable outcomes
Outcome fees start only after an accepted meeting reaches its scheduled time. Never on a reply, a click, or an attendance guess.
Transparency by default
Guardrails, pricing math and sourcing logic are public and inspectable. What “qualified” means is written down before you pay.
The guarantees are in the code, not the marketing.
The same rules that protect recipients also protect our reputation. This is not a story we tell. It is the operating system we run.
If the agent cannot find a verifiable source for a factual claim, the email stops and waits for you. Nobody - not the agent, not autopilot - can release it until the claim is sourced or removed.
Copilot is the default: you approve each batch. Turn on Autopilot and the agent runs on its own inside your daily budget and sending rules. Every send still passes the same guardrail, and a blocked draft never goes out until a person fixes it.
The $25 fee triggers only when a meeting is booked, accepted, not cancelled, and the start time has passed. We refused to charge on “attended” because calendar APIs do not expose it - exactly the kind of unverifiable metric used to inflate numbers.
We are a small team.
We did not build Rainmaker to compete on features. We built it because the trust gap in AI outbound is real and the tools filling the space are making it worse. Every design decision is a deliberate response to a documented failure in the category.
We are in public beta, building in public. The changelog is public and the dispatch goes out with every release.
Rainmaker's outbound runs on Rainmaker. Every beta customer we have was found, researched, drafted and sent by this product - we were the first ones who had to live with what it sends.
Try the thing we built instead.
Paste your domain, read the buyer segments and drafts it proposes for your market, and judge for yourself.