Managed AI guardrails · Steady ground in a daily-changing AI world

AI changes every day. Your ground rules shouldn’t have to.

New AI models, features, and tools ship almost daily. Leveraging managed AI guardrails keeps your AI tools working for you the way you need them to—across updates and across platforms—while we keep the baseline rules current as the AI landscape shifts. Self-serve in minutes for individuals, or tailored and managed across your entire company.

Personal plans (from $9/mo billed yearly) Business & team plans Talk to Business Sales

In plain words

A living handbook your AI reads before it answers.

The AI landscape moves fast. Tools like ChatGPT, Claude, Copilot, Cursor, and Grok Bot roll out new models, agent capabilities, and connectors almost every week. Without a shared baseline, how your AI behaves can drift from one update—or one tool—to the next, with nothing written down about privacy, accuracy, or when a person needs to check the work.

Think of the handbook you’d give a new hire: here’s how we work, here’s what stays private, and here’s when to double-check with a person. Managed AI guardrails connect that handbook directly to the AI tools you already use, so drafts, research, and actions start from steady ground rules instead of guesswork—and stay consistent even as the underlying models change.

Built to help people get more done. AI works best when it augments you and your staff—taking repetitive drafting, sorting, and summarizing off your plate so your time goes into customers, decisions, and high-value work. Choose Personal (Individual) for a single personal login, or Business for a dedicated company endpoint, company-wide domain sign-in, and custom company rules.

How it works

Two paths: self-serve for individuals, guided for businesses.

  • 1. Choose Personal or Business

    Personal (Individual) subscribers pick Standard ($9.00/mo billed yearly at $108/yr, or $10/mo monthly) or Premium ($16.20/mo billed yearly at $194.40/yr, or $18/mo monthly) and sign up online with a personal email address. Businesses talk with our sales team for guided onboarding and pricing sized to their organization.

  • 2. Connect your dedicated endpoint

    Personal accounts are provisioned right after checkout on our managed tier endpoints—mcp.capitolits.com/standard/mcp for Standard or mcp.capitolits.com/premium/mcp for Premium—locked to your personal email address. Business clients receive a dedicated company endpoint (mcp.capitolits.com/<company>/mcp) with state rules and custom company ground rules.

  • 3. Sign in from your AI tools

    Paste your connection address into a supported AI tool, sign in with the 6-digit verification code sent to your email, and add a one-line standing instruction so your AI loads your ground rules at the start of each task. Your AI tool subscriptions stay yours.

  • 4. Stay current as AI evolves daily

    As new AI capabilities, risks, and tool updates roll out, we keep the managed rule library up to date behind the scenes—so your AI tools keep working the way you need them to. Business clients also receive a monthly allowance for custom company rule updates.

Plans & what’s included

Compare Personal and Business guardrails.

Select Personal (Individual) for instant self-serve checkout with a personal email address, or switch to Business to see organization-wide plans, custom company rules, and guided onboarding.

Looking for company-wide guardrails and custom rules?

Personal plans (automated self-serve)

Built for one person using a personal email address (such as Gmail, iCloud, or Outlook). Instant automated setup, read-only managed rule library kept current as AI tools change, and a 10% discount when billed yearly.

Billing cycle: Yearly billing selected (10% off)

Personal · Self-serve

Personal Standard

Core and everyday AI ground rules to keep your personal AI tools consistent and grounded.

  • Core un-amendable safety, honesty, privacy, and human-accountability ground rules
  • Everyday rules for data sorting, privacy, source-backed accuracy, human review, and scam red flags
  • Single-user OAuth sign-in tied to your personal email on mcp.capitolits.com/standard/mcp
  • Ongoing library updates as AI models and tools change
  • Read-only managed rules (no custom company rules)
$10/mo$9.00 / month Billed $108/yr ($120/yr · Save $12/yr)
Subscribe yearly ($108/yr)

Personal · Self-serve · Full Managed Library

Personal Premium

Everything in Standard, plus deeper rules for sensitive data, connected agent tools, and fraud red flags.

  • Everything in Personal Standard
  • Defense against hidden instructions in emails, web pages, and files
  • Stricter Restricted-data handling, payment/wire fraud red flags, and connected-tool approval tiers
  • Output self-checks, source disagreement flagging, and specialized industry rule packs
  • Single-user OAuth sign-in tied to your personal email on mcp.capitolits.com/premium/mcp
$18/mo$16.20 / month Billed $194.40/yr ($216/yr · Save $21.60/yr)
Subscribe yearly ($194.40/yr)

Personal self-serve checkout

Start your Personal subscription

Enter the personal email address you’ll use to sign in (for example Gmail, iCloud, or Outlook). Need company-domain sign-in or custom rules for a team? .

Supported AI plans for individuals: Adding a custom MCP connector requires an AI plan that supports custom connectors—for example Claude Pro or Max, ChatGPT Plus or Pro, Cursor, or Grok Bot. Free-tier consumer AI accounts generally don’t support custom MCP connectors, and Microsoft 365 Copilot and Gemini Enterprise require a business tenant.

Subjects covered: Personal Standard vs. Personal Premium

Every plan starts from the same locked core ground rules. These are a structured guideline that directs how your AI tools are meant to work as models and features change—people still review the work that matters.

Subject area Personal Standard
$9/mo yr · $10/mo		 Personal Premium
$16.20/mo yr · $18/mo
Core & everyday AI ground rules (in both Personal plans)
Core safety, honesty & accountability baseline Lawful use, keeping passwords and private data in place, human accountability for real-world decisions, and locked core rules that cannot be switched off. Included Included
Source-backed answers, no guessing & draft labeling Facts come from sources the AI can name, missing details prompt a question instead of a guess, drafts and estimates are labeled, and intellectual property is respected. Included Included
Human confirmation before deleting data, spending money, or changing settings Explicit human confirmation required for each permanent deletion, financial action, or security/configuration change. Included Included
Privacy, personal details & everyday information sorting Sorting work into Public, Internal, and Confidential levels, minimizing personal details, and keeping sensitive material out of uncleared tools. Included Included
Human review checkpoints & external message boundaries Clear checkpoints before external messages, customer replies, published content, or financial drafts go out, with the AI highlighting what to verify. Included Included
Everyday security basics, approved tools & common scam warning signs Warnings on unknown links, attachments, urgent payment requests, or changed bank details, plus practical record-keeping habits. Included Included
Advanced scrutiny & connected-tool controls (Personal Premium)
Stricter Restricted-data handling (health, banking, government IDs) A fourth, stricter Restricted data tier with explicit stop rules if highly sensitive records appear in the wrong tool. — Included in Premium
Defense against hidden instructions inside emails, web pages, and files Stricter handling when untrusted documents or pages try to redirect the AI, hide text, disguise links, or slip weak spots into code. — Included in Premium
Action controls & step-by-step approval tiers for AI agents and connected tools Read-only defaults for connected tools, sorting every action into read, draft, change, or high-risk before anything runs. — Included in Premium
Stronger financial, payment, wire, invoice & anti-fraud controls Red-flag checks on new bank details, wires, gift cards, mismatched invoices, unfamiliar vendors, and amounts just below approval thresholds. — Included in Premium
Identity & social-engineering checks (executive-impersonation red flags) Messages claiming authority (“this is the CEO”) or pressing for speed and secrecy trigger verification through a separate known channel. — Included in Premium
Output scanning & multi-step verification for high-stakes work Pre-response self-check to mask stray credentials or personal data, plus explicit flagging when sources disagree or confidence is low. — Included in Premium
Connector governance, audit logging, incident response & industry rule packs Structured rules for managing connected integrations, logging high-risk steps, handling AI incidents, and specialized industry rule packs. — Included in Premium

Personal connection endpoints: Personal Standard connects at mcp.capitolits.com/standard/mcp and Personal Premium connects at mcp.capitolits.com/premium/mcp. Both are locked to your single personal email address and deliver our read-only managed rule library.

Need company-wide rules or custom policies? to see dedicated company endpoints, domain-wide sign-in, state rules, and custom company rule layers.

Data and security

Your data doesn’t come to us.

“Where does my data go?” is the right first question. Here’s the plain answer.

  • It’s read-only. Your AI tool reads your ground rules from us. Nothing you or your team types into your AI is sent back to us.
  • We keep a minimal record. Our logs note which rule set was read, never what you asked the AI.
  • Isolated access. Personal sign-ins are locked to your exact personal email address on mcp.capitolits.com/standard/mcp or mcp.capitolits.com/premium/mcp. Business sign-ins are locked to your company’s verified email domain and dedicated path (mcp.capitolits.com/<company>/mcp) so one client can never see another client’s custom rules.
  • We don’t watch your chats. There’s no monitoring of your AI conversations, and we stay out of your chat history.
  • Healthcare and other regulated work: Business Healthcare (HIPAA) adds fixed PHI, BAA, and minimum-necessary ground rules for healthcare teams. This service directs your AI tools and supports your internal safeguards, and does not replace your own legal or regulatory review.

Guideline, not guarantee. Managed AI guardrails are a structured guideline that directs how your AI tools are meant to operate. AI can still misread or skip instructions, so you and your team keep reviewing the work that matters, especially anything going to a customer or into your books.

Q&A

Common questions.

  • Why do I need managed AI guardrails if I already use ChatGPT, Claude, or Copilot?

    AI models, features, and agent capabilities change almost daily. A prompt you wrote three months ago—or a rule living in one person’s head—doesn’t automatically carry across tools or keep pace with new risks. Connecting a managed guardrails endpoint gives your AI tools a single, structured handbook to check before every task, while we keep the baseline rules current for you.

  • What’s the difference between Personal (Individual) and Business?

    Personal is automated self-serve for one person using a personal email address, with our read-only managed rule library: Standard is $9.00/mo billed yearly ($108/yr, saving $12/yr) or $10/mo monthly at mcp.capitolits.com/standard/mcp; Premium is $16.20/mo billed yearly ($194.40/yr, saving $21.60/yr) or $18/mo monthly at mcp.capitolits.com/premium/mcp. Business is built for organizations: it includes guided onboarding, a dedicated mcp.capitolits.com/<company>/mcp endpoint, sign-in across your company email domain, state rules, and custom company rules, quoted during onboarding by company size.

  • Which AI tools and plans work with this?

    Any AI tool that supports custom MCP connectors with OAuth sign-in. For individuals, that includes paid plans such as Claude Pro or Max, ChatGPT Plus or Pro, Cursor, and Grok Bot (free consumer tiers generally don’t support custom connectors). For businesses, it also includes business tenants such as ChatGPT Team/Enterprise, Claude Team/Enterprise, Microsoft 365 Copilot, and Gemini Enterprise. These are examples, not the only options.

  • Does the AI check the rules automatically?

    When your AI tool connects, our server tells it to load your effective ground rules at the start of each task, and we give you a one-line standing instruction to paste into your AI app’s custom or project instructions so it checks them consistently.

  • Do you resell AI tools or licenses?

    No. You buy your own AI subscriptions directly from the vendor, and we provide and maintain the guardrails connection.

  • Can a company use a Personal plan for its team?

    No. Personal plans are locked to a single personal email address with active-session limits and no company customization. Teams and company email domains use our Business plans.

  • Will this replace my staff?

    No, and that’s never the goal. We build and manage AI guardrails to augment your staff—helping the people you already have get more done with confidence, so their time goes into customers, follow-up, and judgment calls.

  • Do you only work around Olympia?

    No. This is remote work, wherever you are. On-site help is available around Olympia, Lacey, and Tumwater when it’s useful.

Business inquiries

Setting up AI guardrails for your company?

Send a short note with your company name, website domain, total employee count, and the AI tools your team uses. I’ll reply with high-level answers and set up your free 30-minute video demo and Q&A.

Talk to Business Sales

Prefer email? Write to sales@capitolits.com. Looking for a single personal login instead? See Personal plans.

Remote anywhere; on-site help available in the South Sound area when useful. Existing clients needing help can reach support@capitolits.com.