Pricing

Pricing answers one question: how many people use Lotics each month. Every size has one price, made of two lines, platform and service, and both grow with it. Records and documents are unlimited at every size. Lotics AI credits are included in the platform; if you want more, an AI bundle adds them on top.

Every plan includes Lotics AI credits. Enough to use it, see what it does, and cover the occasional job. If you want to run more of your operation on it, an AI bundle adds credits on top — an optional purchase, not a decision you have to make before you have used it.

Prices are in Vietnamese đồng, with US dollars converted at 25.000₫ to the dollar. Annual prepay is 15% off, billed upfront, on both lines and the AI bundle. Allowances are the same either way.

By size

Annual prices are shown per month so you can compare them; an annual contract is paid upfront for the whole term.

If you don’t need us building, drop the service line: you pay for the platform and build the system yourself with your own AI. The table is a starting point rather than a menu — tell us how your operation runs and we will put you on the right size, or price something else if none of them fit.

Platform

Every workspace sits on a platform tier, which follows the size above. A new account starts on Free to try it.

Invite as many members as you like, on every tier including Free: the price follows the people who use it each month, not the number of accounts. Permissions, roles and audit logs are on every tier as well: gating the controls that keep your data safe would make the cheaper tiers less safe, so we don’t. External AI through the CLI, MCP server and REST API is on every tier and consumes no credits. Enterprise adds integration with your internal systems or other software, SSO, data residency and an SLA.

Apps you publish are served on a Lotics address, on every tier.

Lotics AI bundles

Credits meter work that runs on our models: reading documents, extracting data, drafting text, answering questions, and AI steps inside workflows and app agents. How many depends on the work: a message or a simple action takes about 3–5 credits, heavier work like reading a multi-page document or reconciling about 10–20, and about 5–10 on average — so 750 credits a month covers roughly 75–150 actions.

Your plan already includes credits every month — see the platform table above. That covers trying Lotics AI out and the occasional job. If you want to run more of your operation on it, add a bundle. It stacks on top of what your plan includes, and the bigger the bundle, the better the rate.

Every new account gets a one-time grant of 500 credits, on top of what your plan includes each month. It burns down and never refills: it is there so you can set up your workspace and see what Lotics AI does before the monthly credits take over. One grant per person, not per organization.

Credits are indexed to what a model costs us, so when a model gets cheaper the same work simply consumes fewer credits.

Service

The service line is us doing the work: we walk your operation, build the apps around your documents and rules, migrate your data, train your team, and keep building as your processes change.

It is one service, added to your platform tier as a second line item, and what sets its fee is how many people use the platform each month:

That number is not your headcount: a five-person desk inside a hundred-person company is a five-person band. It is also not a seat charge. The platform never bills per account, on any plan; the service fee follows the number of people we serve, which is how many work on the platform each day. We count it each month and tell your admins when you cross a band; the fee moves at your next renewal, and nothing stops in between. What you get is the same at every band: we build what you name, we operate it, and changes land as fast as we can, usually within the week.

Improving it is part of the fee. We run these desks with you, so we see where the work still drags — a step people route around, a document still keyed by hand, a report someone rebuilds every Monday — and we keep changing the system to take it out. Those changes are free and unlimited, like the ones you ask for.

How deep a build we take on is a scoping judgement we make with you on the discovery call — ask us directly what the service does and doesn’t cover and you get a direct answer.

Allowances, and what happens when one runs out

On every paid tier, records and generated documents are unlimited. How many shipments or orders a month an operation runs varies too widely between trades to price on, so the size is the people who use it, and the one allowance that follows usage is AI credits.

Free is for trying it, so it is capped: 1,000 records and 100 generated documents a month. Records are the live rows in your workspace; archived records don’t count, so to bring the number down you archive rather than delete.

Credits meter work run on our models, as above.

When you reach an allowance, we tell you. AI credits stop there until you add a bundle or the next cycle starts; everything else keeps working. Nothing is billed automatically — there is no surprise on your invoice, ever, on any plan.

Credits are the exception we’ll discuss if you ask, because they are a real cost we pay a model provider — but the bundle is almost always the better answer: it is never priced above the overage rate, and it renews instead of surprising you once.

Changes, and what they cost

On a service tier, we build the process and we operate it. Changing that process is free and unlimited for as long as you’re a customer — no change fee, no re-scoping charge, however many times the rules change.

Processes you build yourself stay free and unlimited on every tier, including Free. The service line is not a per-process fee: what it prices is our time, which is the thing that actually runs out.

There is no setup fee, no implementation fee and no build fee on either line (không có phí thiết lập). Onboarding sits inside the service subscription.

Chargeable work starts where the platform stops: an integration with an outside system, bespoke development, or anything needing a capability we don’t already ship. We scope those with you before any work begins.

Lotics AI vs. external AI

Lotics AI reads your workspace schema directly — the field types, the option lists, the links between tables — so it writes into a record without being told what shape the record is. It also runs inside the platform: an AI step in a workflow, a document generated from a template, an agent embedded in an app are all Lotics AI, and none of them are reachable from outside.

Your own model, connected over the CLI, SDK, REST API or MCP server, works inside your workspace. It consumes no credits, because the model never runs on our tokens. What it cannot do is run as a workflow step or as an app agent — those execute on our side, on our models, and are metered in credits.

Bring your own AI

You can also connect your own model — ChatGPT, Gemini, Claude, or any other — to Lotics through the CLI, SDK, or REST API, on every tier including Free. They consume no credits, because the model never runs on our tokens.

Running your own model may mean you never need an AI bundle; it rarely means you use no credits at all. Three things keep drawing them:

• Teammates without a subscription. The person who built the tool pays for a model; the eight people on the ops desk usually don’t, and eight separate subscriptions cost more than one bundle the whole desk draws on.
• Occasional work. Nobody keeps a subscription alive for the fortnightly job.
• Agents inside your apps. An agent built into an app runs on our models by design and is metered in credits, whatever you use elsewhere.

Cancellation and refunds

• Monthly plans: Cancel anytime. No refund for partial months. Access continues until the end of the current billing cycle.
• Annual plans: The term runs for the 12 cycles you paid for, and does not renew automatically — we contact you 30 days before it ends with the price for the next term, and it ends unless you confirm. Outside the two windows below, prepaid fees aren’t refundable and access continues to the end of the term — unless we seriously breach our service commitments, in which case the unused remainder of the term is refunded. Allowances stay active throughout.
• 14-day refund window: Full refund within 14 days of the first payment for all new annual subscriptions.
• First-cycle exit (annual): If the processes we build aren’t right for you, you can end the contract at the close of the first cycle by telling us within 10 working days of going live. You pay for that cycle only, and we refund the rest of the prepayment within 15 days. This applies even after the 14-day window has closed.
• Data export: Available at any time on all plans, including after cancellation during the remaining access period.

Billing runs from your signup date rather than the calendar month, so signing on the 28th bills 28th to 28th. There is no proration in either direction — the cycle already starts when you do.

The discovery call is free and commits you to nothing. All data exportable at any time.