Rallly

What Rallly is, and what it is not

The product is deliberately small. You create a poll with several date and time options, send the link, and each participant marks which options work. An availability grid shows everyone's responses together, comments let people argue about the details in one place instead of five reply-alls, notifications tell you when responses land, and when you pick a winner everyone gets told. The feature that carries most of the weight is the one that sounds least impressive: participants do not need an account to vote.

It is TypeScript on Next.js, Prisma and tRPC, licensed AGPL-3.0, published as a Docker image with self-hosting documentation, and interface translations are crowd-sourced across more than ten languages. A hosted version exists, run by the same maintainer, which is the sensible default unless you have a specific reason to hold the data yourself. What it is not is a booking page: it does not publish your calendar so a stranger can claim a slot. If your problem is inbound demo requests, you want that other category of tool — Documenso sits at the next step of the same deal, once a time has been agreed and there is something to sign.

Where group scheduling actually costs a founder time

The place this earns its keep is customer research, not sales. A round of user interviews is a scheduling problem disguised as a research problem, and it is where founders quietly lose a week.

1. Recruit more people than you need Ask ten users for a conversation when you want six. Reply rates to a research request are decent, attendance rates are not, and over-recruiting is cheaper than a second recruiting round.
2. Offer options, not a calendar Send one link with six or eight slots across two weeks rather than asking what suits. An open question generates a thread; a set of options generates an answer. This is the step Rallly exists for, and it is the one people skip.
3. Lock the times and confirm in writing Finalise as soon as the votes are in and send the confirmation the same day. Time between agreeing and confirming is where a research call quietly evaporates.
4. Write the notes before the next call Scheduling six conversations is only useful if you can compare them afterwards. Ten minutes of notes immediately after each one beats a transcript you never reread — see how to collect customer feedback for what to do with the material.

An honest note on which scheduling problem you have

We build AgentCeres — the AI Growth Officer at agentceres.com — and the useful observation from operating it is that these two scheduling jobs get confused constantly, in the tools people adopt and in the advice they follow. One is many-to-one: several people, one meeting, availability unknown, which is a poll. The other is one-to-many: your availability published, strangers picking a slot, which is a booking page. Adopting the poll for the booking job produces a worse funnel than plain email, because you have inserted a voting round in front of somebody who was ready to buy.

The rule of thumb we use: if the person on the other end has already decided they want to talk to you, never make them vote. If nobody has decided anything yet — a research round, an advisory call, a partner intro across three companies — the poll is the faster instrument, and one AGPL-licensed self-hostable one is a reasonable thing to run. Either way it is a small tool doing a small job well, which is more than most of the category manages. If the calls it books are sales calls, the harder part is upstream in founder-led sales, not in the calendar.