← Go back

Guide // MandateGuard

Free vs Pro vs Team: which MandateGuard plan fits your agents?

By the Ingenious AI team · Published 2026-11-14 · Reviewed 2026-11-14

▸ Run the free live demo See pricing

Free covers 1 agent and 1 policy with the full authorize flow, signed receipts and the approval inbox — no card required to start.

Short answer

Free runs MandateGuard's complete control loop for one agent — authorize decisions, signed receipts, approval inbox and webhook — so it is a real deployment, not a demo. Pro (29 dollars a month) adds the scale and depth a growing fleet needs: multiple agents and policies, vendor trust from your own ledger, duplicate-commitment detection, the kill-switch, evidence export, MCP tools and the interoperable intent service. Team (79 dollars a month) is Pro with higher limits and priority support. Upgrade on triggers evidenced by your own receipts — a second agent, a duplicate, an audit request — not on feature envy.

What do Pro and Team add over Free?

Free, Pro and Team are MandateGuard's plan tiers. Free: one agent, one policy, the full authorize flow, signed receipts, approval inbox and approval webhook — the complete control loop for a single agent. Pro (29 dollars a month): multiple agents and policies, vendor trust from your own ledger, duplicate-commitment detection, kill-switch, evidence export, MCP tools and the interoperable intent service. Team (79 dollars a month): everything in Pro with higher agent and policy limits and priority support.

What Free actually includes — and why it matters

Free is the complete control loop for one agent: workspace, agent, one policy with caps, allowlist and threshold; the full authorize flow with all three decisions; signed hash-chained receipts with verification; the approval inbox; and the approval webhook. That completeness is a deliberate product stance — the free tier does the whole job for the single-agent case — and it changes the buying conversation. You are not evaluating a demo of control; you are running control, and the paid question becomes purely about scale and depth: more agents, cross-agent detection, evidence packaging, shared mandates. Teams should exploit this: put the free plan on the agent whose spend worries you most, and let a quarter of receipts, not a sales page, size the paid case.

The paid depth, item by item

Pro's additions each answer a specific scale pain. Multiple agents and policies: the second job gets its own fitted mandate instead of sharing or waiting. Vendor trust: per-vendor behaviour from your ledger, for procurement conversations. Duplicate-commitment detection: the cross-request waste class, escalated before it completes. Kill-switch: one-action containment, the difference between an incident and a scare. Evidence export: audit-ready packs for finance and external reviewers. MCP tools: the policy check inside the agent's native tool loop. Interoperable intent service: mandates shared with merchants and issuers, with role-scoped verification. None of these is decorative; each maps to an incident or a meeting that single-agent Free cannot serve.

Team: limits and support, honestly

Team is Pro's feature set with higher agent and policy limits and priority support, at 79 dollars a month. That is the whole difference, and it should be bought for exactly those reasons: your counts approach Pro's ceilings, or your operation's criticality justifies priority support. Buying Team early for ambition's sake mainly buys headroom you cannot yet use; buying it late shows up as friction at the moment a new agent cannot be registered. Watch the counts at the monthly review — agent and policy totals are displayed facts, not estimates — and let them, plus any client or audit requirement for support, time the step.

The trigger list for upgrading

Concrete triggers beat calendar-based upgrades. From Free to Pro: a second agent or policy is needed; a duplicate commitment occurs (or nearly does); an incident makes the kill-switch's absence felt; an auditor, client or insurer asks for evidence in pack form; a counterparty mandate needs the intent service; agents move to MCP-native frameworks. From Pro to Team: agent or policy counts press the limits; support criticality rises (client-facing fleets, contractual SLAs). Notice every trigger is an event in your operation, evidenced by receipts or requirements — which means the upgrade conversation can be conducted entirely from your own records, the healthiest way to buy anything.

What does not change across tiers

Three things are constant, and they are the ones that matter most. The decision engine is the same: ALLOW, DENY and REQUIRE_APPROVAL mean the same on every tier, evaluated against the same four policy parts. The receipt chain is the same: hash-linked, signed, verifiable, continuous across upgrades — your free history becomes your paid history without migration or loss. And the method is the same: fit caps to real prices, allowlist deliberately, staff the inbox, review monthly. Tiers change how many agents and how much depth; they never change what good looks like. A team running Free well is better governed than a team running Team carelessly, and the receipts will say so in either case.

A note on bundles

MandateGuard is also sold inside the fleet's bundles — the Security and Safety Bundle and the Agent Safety Bundle, plus the Everything Fleet Pass — through the same PayPal door, with the pricing page showing the current bundle composition and math. Bundles make sense when your agent programme is part of a wider Ingenious AI deployment (identity in Agent Passport, receipts in Agentic Receipts, and so on). The tier logic in this guide still applies inside a bundle: choose by agent counts and the paid-depth triggers, and let the bundle be a pricing vehicle rather than a reason to adopt platforms you have not evaluated. Control software should be bought the way it teaches you to buy everything: on evidence, at the boundary, with the receipts in hand.

Five actions to take now

  1. Start free with your highest-risk single agent, not the least important one
  2. Upgrade when a second agent arrives or the first cross-agent duplicate does
  3. Count the free plan's receipts as an asset — they seed trust and audits later
  4. Do not buy Team for limits you do not need; counts, not ambition, decide
  5. Review the trigger list quarterly as the fleet changes

Worked example: upgrading on evidence, twice deferred

A consultancy runs its enrichment agent on Free for a quarter: 1 agent, 1 policy, full receipts, approvals in the inbox. The quarter's only gap appears in month three, when a second workflow (run by the same agent identity, a setup smell the team already disliked) double-buys a dataset — visible in receipts, caught by a human, but manually. The upgrade decision writes itself: a second agent for the second workflow, plus duplicate detection so the next overlap is escalated automatically. Pro at 29 dollars a month against a demonstrated 300-dollar-a-quarter duplication pattern is not a hard meeting.

They stay on Pro for a year. Team enters the conversation only when the agent count and a client's audit requirement (priority support, higher limits) make it concrete. The tier history — Free until the evidence demanded Pro, Pro until counts demand Team — is the honest shape of buying control software: each step justified by receipts, not by feature envy.

Frequently asked questions

Is Free a trial with features missing from the core loop?

No. Free runs the complete control loop for one agent: authorize decisions, receipts, approvals, webhook. Paid adds scale and depth around that loop.

What is the single most valuable paid feature?

For most fleets, duplicate-commitment detection — it catches the waste class that no per-request control can see. For counterparty work, the intent service.

Do my receipts carry over when I upgrade?

Yes. Your ledger is continuous; vendor trust computes from the history you already built.

How are paid plans bought?

Through the pricing page: PayPal door today (license key emailed after confirmed payment), with crypto checkout via BTCPay where live. Keys activate the plan on your workspace.

Can I run production agents on Free?

One agent, genuinely in production, yes — many teams do. The limits that bite are agent count and the absence of duplicate detection, kill-switch and export, not decision quality.

Keep reading

Put the policy in force: spend envelopes · intent registry · approval inbox · how the proof works · all MandateGuard guides