Dot
Dot is the agent in Limitr Cloud. It works with you inside the app. Ask it in plain words: "add a 14-day trial to Pro", "which customers are over their limit this month", "show me margin by model for the last two weeks". It reads what's on your screen (the page, the workspace, test or live, and the segment filter), so you rarely have to explain where you are.
Open it from the panel on the right of any page, or from anywhere with Ctrl K (Cmd K on a Mac). Dot in the left rail is the full-width version, with your past conversations. Conversations are yours: they're private to you and kept per workspace.
What it can do
- Write your policy. Dot changes the workspace's draft, the same draft you and your teammates edit, and says what it changed. Review it on the Policy page (the Changes tab shows exactly what moved).
- Check its work. Before it tells you a policy does something, it runs made-up customers through it with simulate and tells you what it checked. Nothing touches your real customers.
- Read your data. Customers, usage, events, and segments, to answer questions about them. It filters by the segment you have selected unless you ask about everyone.
- Show you. When a picture answers better, it shows a view in the conversation (usage over time, cost by model, margin per task). Open it full size, or save it as a dashboard (Enterprise). See Dashboards.
- Propose changes for you to apply (below): to customers, segments, notifications, and what's published.
- Take you there. Ask to go somewhere ("open Acme", "show me my billing") and the app takes you there, with the conversation still open beside the page. Otherwise it offers a button.
- Answer how-to questions from these docs, with a link to the page.
- Look things up on the web, like a competitor's pricing page or a provider's token prices, and say where the answer came from.
Changes you apply
Apart from your draft, Dot never changes anything itself. When you ask for a change ("move Acme to Enterprise and give them 500,000 chat tokens until December"), it proposes it on a card in the conversation, with each change's before and after. Nothing changes until you press Apply. Choose Discard to drop it.
Dot can propose:
- Customers: adding one; their label, type, metadata, and refs; their plan; giving, taking, or restoring an entitlement; a limit of their own; caps; a topup, free of charge; and their own prices (Enterprise).
- Segments and notifications: making, changing, and deleting them.
- Publishing: publishing the draft in the mode you're in, with what changes from the live version, or copying test mode's policy into the live draft. Publishing the live draft is always a card of its own.
When you apply a card:
- It applies as you, with your permissions, checked again when you press Apply. A card in live mode says so, and its button reads Apply to live.
- It applies only what it showed. If something a change depends on moved since Dot proposed it (a teammate changed the customer's plan, or saved the draft), nothing on the card is applied, and the card shows which change is out of date. Ask Dot to look again.
- It's recorded: history and the event log show "Dot for" your name.
- A card can be applied for a day, and Dot hears whether you applied or discarded it.
Dot can't delete customers, charge anyone, change how a customer is billed, or touch API keys, the team, Stripe, or integrations, and it never sees secrets. When you ask for one of those, it tells you where in the app to do it.
By plan
| Free | Pay as you go | Enterprise | |
|---|---|---|---|
| Writes and simulates the draft, answers questions, navigates | Yes | Yes | Yes |
| Reads customers, usage, and events; proposes changes | Yes | Yes | |
| Web search | Yes | Yes | |
| Views in the conversation | Yes | Yes | |
| Saves dashboards | Yes | ||
| Notes to the Limitr team | Yes | Yes | |
| Model | The standard model | The standard model | The largest model for deep work (policy design, analysis), the standard model for the rest |
Usage
Dot's usage is what its models cost. The line under the composer shows where you are:
- Free: a small amount to try it, once ("$0.62 left to try Dot"). When it's used up, Dot shows the plans; everything you made stays.
- Pay as you go: an amount included each month ("$3.42 of $5 included this month"), then usage is billed at cost plus a small markup. See Plans and billing.
- Enterprise: usage is included in your contract (Settings > Plan shows it).
A very long conversation eventually can't take more turns. Start a new one; Dot can read your policy and data again in a moment.
Notes to the Limitr team
When you need something Limitr can't do yet, or something that should work doesn't, Dot offers (once) to tell us. If you say yes, it writes a short note: what you wanted or what happened, what you tried, and anything that helps us reproduce it. Nothing is sent until you press Send to the Limitr team. You can add a note of your own first, or choose Don't send.
The note goes to support@limitr.dev with your name, your email (so we can reply to you directly), the workspace and its plan, and a reference to the conversation. We read every one. You can send a few a day; for anything more, email support@limitr.dev.