Every plan is the same bot.
What a bigger plan buys is how often it is allowed to think — and thinking is the only thing that costs money. Watching your group is free on every tier.
Prove it works in your own group. 300 messages or 14 days, whichever runs out first. No card.
- 300 messages analyzed / cycle
- No overage
- 1 chat group
- 1 team seat
One group, one niche. Proof that it works.
- 1,000 messages analyzed / cycle
- then $0.025 per message
- 1 chat group
- 3 team seats
The plan the business runs on. Multi-group, with headroom.
- 3,500 messages analyzed / cycle
- then $0.018 per message
- 3 chat groups
- 5 team seats
Large communities, the deepest volumes.
- 10,000 messages analyzed / cycle
- then $0.015 per message
- 10 chat groups
- 10 team seats
Running something bigger?
More groups, higher volume, or terms the standard plans don’t cover — talk to us about custom pricing.
You are billed for messages the assistant thinks about
Not every line typed in your group — only the ones sent to the model for a decision. Watching the room is free and unlimited, the replies it posts are included, and members are never counted.
Under your limit
Nothing to watch. The bot reads the room and answers when it has something worth saying.
Close to your limit
The dashboard shows what is left and what going over would cost. The bot carries on as normal.
Over your limit
Extra messages bill at your plan rate and the bot carries on. The trial is the exception: it has no overage, so it stops. Recommendations are metered separately, with their own monthly cap.
7 days to fix a failed card
The bot keeps running for 7 days while Stripe retries, so an expired card is an email rather than an outage.
Cancel whenever
Canceling stops the next charge. You keep the plan you already paid for until the cycle ends.
The commission is yours
Your affiliate tag, your earnings. The subscription is the whole of what we charge, and we never touch the money.
Every limit, side by side.
The rows below are the plan itself — each one is a row in the database the bot reads, not a promise on a marketing page.
Cycle: A cycle is one billing period. It starts the day you subscribe and resets on each renewal, so your usage goes back to zero every time you are charged. The free trial has a single cycle: 14 days from signup.
| Limit | Free trialFree | Starter$20/mo | Growth$59/mo | Scale$149/mo |
|---|---|---|---|---|
| Included messagesPer cycle — messages the assistant reads closely enough to decide on. Watching the room is free and unlimited; only the messages sent to the model count, which is the thing that actually costs money. | 300 | 1,000 | 3,500 | 10,000 |
| Then per messageCharged for each analyzed message past the included volume. Every tier prices overage above the next tier up, so upgrading always costs less than overrunning. | No overage | $0.025 | $0.018 | $0.015 |
| Product recommendationsPer cycle — each product the bot posts counts once. Clarifying questions, replies and declines are free; the recommendation is the unit because it is the thing that can earn a click. A quality cap, not a billing one. | 25 | 300 | 1,000 | 3,000 |
| Chat groupsHow many Telegram or Discord groups the bot runs in. | 1 | 1 | 3 | 10 |
| Team seatsPeople who can sign in and manage the bot. | 1 | 3 | 5 | 10 |
| Group messages watchedUnlimited on every plan. Watching the room is free — the meter only counts messages the assistant reads closely enough to act on, a small fraction of traffic. | Unlimited | Unlimited | Unlimited | Unlimited |
| Included rateWhat one included message works out at. Compare it with the overage row: every plan bills overage above the next plan’s included rate, so upgrading always beats overrunning. | — | $0.020 | $0.017 | $0.015 |
What your plan changes about the bot.
It reads everything, always
Watching the room is free on every plan and is never rationed. The bot is not mention-gated — joining in when it has something useful is the product, not a feature of the expensive tier.
It waits for you to finish
People say what they want across several messages — “I need a speaker”, “for my room”, “under $50”. The bot waits for a pause before spending a call, so it answers the whole request instead of the first fragment. Being @mentioned skips the wait.
Speaking is capped for quality, not billing
Replies are included in the price. How often the bot may start a conversation uninvited is yours to set — the same control on every plan, there so it never becomes noise, not a meter.
A bigger plan is a more attentive bot
More included messages means a higher daily budget, so in a busy group it evaluates more often and catches more of what passes. In a quiet group the cheapest plan already sees everything worth seeing.
Try it in your own group.
300 analyzed messages or two weeks, whichever runs out first, with no card. Enough to tell a working bot from a broken one before you decide anything.