Town's AI Assistant Runs on Credits — Here's What That Actually Costs You
Short answer: Town prices its AI assistant on a credit system — a free tier capped at roughly 30 chats a month, then paid plans reported to run from about $15 up to $199 a month depending on how many credits you buy, plus pay-as-you-go overage at $0.03–$0.044 per credit once you run out. There's no single flat number, because your bill depends on how much you actually use it in a given month. That's worth comparing carefully against bots.team, and fairly: Town's price includes the AI itself, while bots.team's $0–$19/month tiers price the app only — you bring your own Claude subscription (from $20/month) to actually power your bots, so the real bots.team total is closer to $20–$39/month, not $0–$19 alone.
Town's name evokes something local and familiar — its pricing model, however, is closer to a metered utility bill.
What is Town, and why does it matter to bots.team's audience?
Town is a personal AI assistant that raised a $55 million Series A in June 2026, led by Andreessen Horowitz with Forerunner Ventures also participating. What makes it worth watching isn't just the funding — it's who Town says it's for. In its own announcement, a16z described Town as an assistant "that learns how they work, not one that requires them to learn how it works," and named the kind of people it's built around: nonprofit directors, parents, teachers, founders — not engineers, not power users.
That's the same non-technical audience bots.team is built for. You give Town a "Townie" — its own email address and a presence on Slack, WhatsApp, or Telegram — forward it work, and it acts across roughly 50 connected tools. It ships pre-built routines out of the box: morning briefings, inbox triage, contact research, newsletter digests. Mechanically, that's close in spirit to how bots.team works: describe a job once, let it run without you re-prompting it every time.
Where the two products genuinely differ is pricing structure — and that difference is worth understanding before you pick either one.
How does Town's credit pricing actually work?
Town's own pricing page returned an access error when checked directly, so exact current tier numbers should be treated as reported rather than officially confirmed — a caveat worth keeping in mind before quoting a specific dollar figure as gospel. That said, multiple independent reviews published in mid-2026 converge on the same shape:
- 30 chats/month on the free tier
- $15–$199 reported range of paid monthly tiers
- $0.03–$0.044 per-credit overage once you exceed your plan
The mechanism is straightforward: each plan comes with a monthly credit allotment. Chats, routines, and tool actions each consume credits at some rate. Run out before the month resets, and you're billed per credit at the overage rate — no cap, no warning built into the sticker price. One review summarized the free tier bluntly as "more of a preview rather than a usable daily tool," since 30 chats a month works out to roughly one a day — thin for something meant to run recurring routines on your behalf.
Why this is worth checking before you sign up: a credit meter isn't a bad pricing model on its own — it lets a provider charge roughly in proportion to compute cost, which is fair in principle. But it also means the number on the pricing page isn't the number on your card statement. If your assistant runs a daily morning briefing, triages your inbox, and researches a contact or two, credits add up in ways that are hard to predict until you've already used the product for a month.
What "credit-metered" pricing tends to cost in practice
This isn't unique to Town — it's the same shape Microsoft uses for its enterprise agent product, MS Copilot Cowork, where a $30/month seat license includes zero task execution and every task is billed separately at $0.01 per credit (a "heavy" task can run $7 or more on its own). The pattern shows up whenever a product wants granular cost attribution. It's a reasonable choice for an IT department billing usage back to departments. It's a much harder thing for an individual to budget around, because you don't find out what a busy month actually cost until the bill arrives.
Credit-based pricing means your bill scales with how much you actually use the assistant that month — good for occasional use, harder to predict for daily, recurring work.
Town vs. bots.team: two different pricing philosophies
Both products are chasing the same idea — a non-technical person should be able to describe a job in plain English and have an AI helper carry it out, repeatedly, without being re-prompted each time. Where they split is how they charge for it, what they show you when the work is done, and what they remember.
| Question | Town | bots.team |
|---|---|---|
| Pricing shape | Credit-metered, AI included — reported $15–$199/mo tiers + pay-as-you-go overage | App is flat (Free $0, Basic $5/mo, Pro $19/mo) but doesn't include AI — add a Claude subscription (Claude Pro from $20/mo, also flat) |
| Free tier | ~30 chats/month (light preview use) | Unlimited bots and unlimited runs on the app; still need a Claude subscription to run them |
| Cost predictability | Varies month to month with usage; overage has no hard cap | Two flat numbers (app + Claude Pro) as long as usage stays within Claude Pro's limits |
| Reach | Email address, Slack, WhatsApp, Telegram, ~50 connected tools | Desktop app (Mac and Windows): scheduled/triggered bots working on your files and the web |
| Action safety | Three modes: read-only, approval-required, autonomous within set parameters | Bots run on an approved standing plan you set up once — a deliberate design choice, not an unfinished feature |
| Where results show up | Messages sent to your email address, Slack, WhatsApp, or Telegram — one channel per message | An organized dashboard: reports, a discovery feed, and metric charts, not just a stream of messages |
| What it remembers | Context within a routine's own conversation history | Each bot keeps its own memory across runs, plus a shared team memory other bots can draw on |
| A busy month for a daily-use case | Can push well past the base tier via overage — no built-in ceiling | Same $5–$39 or $19–$39 all-in, whether a bot runs once a day or every hour, within Claude Pro's usage limits |
Worth saying plainly: Town bundles the AI into its price. bots.team doesn't — it's a wrapper around your own Claude account, so its app fee alone understates what you'll actually pay.
Neither structure is objectively wrong. If you'll only lean on an assistant occasionally, a credit meter can genuinely cost less than a flat subscription. But if you're hiring an AI helper to do recurring work — the exact use case both products are selling — a flat rate means you can commit to a routine without doing mental math about whether this week's usage is going to blow past the plan.
What Town gets right that's worth acknowledging
It would be easy to write this off as "usage-based pricing is worse," but that's not quite fair. One detail worth crediting: independent reviews describe Town shipping three distinct action-safety modes — read-only (it only observes), approval-required (it proposes an action and waits for you to confirm), and autonomous (it acts on its own within limits you set). That's a real, publicly documented safeguard for a product that's reaching into your email, Slack, and messaging apps on your behalf. It's a genuinely useful design decision, and one that any non-technical user evaluating an AI assistant should look for — regardless of which product they pick.
An approval-required mode — where the assistant proposes an action and waits for a yes — is one of the clearest signs a tool takes autonomous-action risk seriously.
Questions to ask before picking any personal AI assistant
Is the price on the homepage the price I'll actually pay? Not necessarily, if the product is credit-metered. Ask specifically what a "typical week" of your intended use would cost, not just what the entry tier includes — and ask what happens the moment you exceed it (hard stop, or automatic overage billing).
Does it ask before it acts, or does it just act? Look for an explicit approval mode, not just a settings toggle buried in documentation. Town's three-tier model (read-only / approval-required / autonomous) is a good example of what to look for in any tool that can touch your inbox, calendar, or files.
Is Town available yet, and is it a fit for individuals vs. teams? Yes — Town is live with an iOS app and web access as of mid-2026, still early (its App Store listing had a small number of ratings as of this writing). It's positioned for individuals first, which puts it in more direct competition with bots.team than most of the workplace-productivity tools in this category.
How does bots.team price its equivalent features? Flat and tier-based, not usage-metered: Free ($0, unlimited bots and runs, no account needed), Basic ($5/month, adds more team areas and scheduling options), and Pro ($19/month, unlimited team areas and advanced scheduling). That's the app price, though — bots.team doesn't supply the AI itself, so you'll also need your own Claude subscription (Claude Pro from $20/month) to actually run bots. All-in that's roughly $20–$39/month, still two flat numbers you can know in advance rather than a credit meter.
Bottom line: if you want an assistant that shows up in the channels you already check, Town's spread — email, Slack, WhatsApp, Telegram — is genuinely broader than bots.team's desktop app. But if you want to know exactly what a busy month will cost before it happens, and you want your bots' work organized into reports and dashboards you can scan at a glance rather than dug out of a message thread, bots.team's flat two-number pricing and standing per-bot memory are the sharper fit.