How do I…

The jobs teams use Limitr for. Each answer says how Limitr does it and links to the doc. All of them run on one policy and the same engine in your app, so they combine: a budget, a contract, and a rate limit can all apply to the same call.

Cap AI spend per user or per org

Put a spend cap on the customer, in dollars or credits, and share it across an org if you like. Limitr checks it in your app before the call, so the call that would go over never runs and you never pay the vendor for it. Warn customers before they reach it with a notification.

Docs: Budgets and spend limits · Solution: Agent spend controls

Give each department its own AI allowance

Make departments, teams, and employees customers: give each a type (department, employee) and refs from an employee to their department and from a department to the org. Put each department on a plan, or give it a shared monthly cap, and give employees their own caps inside it, so an employee is stopped by whichever runs out first. IT or finance changes any allowance live in Limitr Cloud, and usage shows by department, team, and employee for showback or chargeback.

Docs: Budgets for an org and its teams · Customers and refs

Cap AI spend in a tool you built for yourself or a few people

Add the open-source engine to your app and give each person a monthly cap in dollars, across every provider you call. It runs in your process with no account. The quickest way is to ask your coding agent: Limitr ships with skills, so it knows the policy and the SDK. If it's only you on one provider, that provider's own limits are enough.

Docs: Quickstart · Budgets and spend limits

Bill customers for AI tokens, with margin

Give each model a credit with your price and your cost per input and output token. Every call is metered at both, so invoices match what customers used and margin per customer, feature, and model is live.

Docs: Price AI tokens per model · Solution: Usage-based billing

Give an AI agent a budget

Make the agent a customer with the agent type and its own plan, link it to the workspace it works for, and open a budget for each task. Every step counts against the task, the agent, and the workspace at once, and the agent can see what it has left as a Claude tool.

Docs: Reserve and settle LLM calls, AI tools · Solution: Agent spend controls

Stop abuse of a free tier or a leaked key

Add burst and steady rate limits per user, daily caps per tier, a budget per org, and a global backstop over every customer. They're checked in your app before the model runs, and you can pause one key without touching anyone else.

Docs: Rate limits next to paid usage · Solution: Security and abuse prevention

Ship an AI feature to paid plans only, then roll it out

Put the feature in the plans that include it, roll it out to a percentage of them, and keep a kill switch. Because the flag lives in the policy that prices the feature, every use is metered and billed too.

Docs: Feature flags, rollouts, and experiments · Solution: Feature flags and entitlements

Enforce a custom enterprise contract in the product

Write the deal as a contract in Limitr Cloud: prices, limits, features, caps, and dates for one customer, an org, or a segment. It reaches every engine on the next call, with no plan per deal and no code.

Docs: Trials and custom deals, Contracts in Limitr Cloud · Solution: Custom contracts

Know which customers, features, and models make money

Every call carries what you charged and what it cost you, so margin is computed as usage happens, by account, feature, and model. Ask Limitr Analyst where the margin is and save the views your team needs.

Docs: Cost and margin, Dashboards · Solution: Usage intelligence

Test a price change before it ships

Ask Limitr Analyst to replay the change on your real usage. You see revenue, cost, and margin per account before you publish, then publish to every engine with no release.

Docs: Limitr Analyst

Move from flat subscriptions to usage-based or hybrid pricing

Map your pricing page to a policy line by line, run it in observe mode next to what you have, and switch enforcement on when the numbers match. Each feature can have its own pricing model.

Docs: Map your pricing model to a policy, Change plans

Sell prepaid credits and credit packs

Sell credits that every feature spends at its own rate, with monthly included credits, packs on top, and features that stop when credits run out.

Docs: Prepaid credits and credit packs

Price seats and usage together

Count seats on the org whichever user invites someone, include some, bill extra, and put usage pricing next to it in the same plan.

Docs: Seat-based pricing on the org

Predict an LLM call's cost before it runs

Limitr learns each model's output from its input and predicts the next call's tokens, so you can reserve the right amount, set max_tokens, quote a price, or pick a cheaper model.

Docs: Predict LLM output tokens, Reserve and settle LLM calls

Set moderation thresholds, allowed models, and strikes per plan

Keep safety settings in the policy as entitlement fields, meter flagged outputs as strikes against the account, and change any of it live without a release.

Solution: Guardrails and governance

Control a network of agents

Connect your agents, your customers' agents, and your partners' in a private network on Limitr Network. Every agent gets one budget across every provider it uses, every call carries a signed receipt, either side can revoke a connection at once, and spend shows by agent and by provider in Limitr Cloud. Nothing is billed unless a provider prices its calls.

Let other apps and AI agents buy from your app

List your plans on Limitr Network, privately to the workspaces you name or publicly on Limitr Cloud. Buyers sign a voucher for a plan, an amount, and a time instead of using an API key; each buyer is just another customer in your policy, and if you price your calls you're paid through Stripe.

Page: Limitr Network

Still deciding?

Is Limitr right for you? · Limitr answers · Every comparison