# Everything your gateway does

One gateway for every AI request your company makes, with what you need to run AI like any other part of the business.

## See where every AI dollar goes

- **Spend by team, person and project.** Every AI request has an owner.
- **Spend by AI company and model.** Compare what each one really costs you.
- **Charge AI back to cost centres.** Every request is priced and attributed, so finance can charge AI back like any other cloud cost.
- **Every request on record.** Who asked, which model, what it cost, how it went.
- **A needs-attention list.** Unusual spend surfaces on its own.
- **Daily spikes and biggest movers.** Spot a runaway agent the same day.
- **Month-end forecast.** Know where spend will land before the invoice does.
- **A personal view for every employee.** People see their own usage and budget.
- **Which AI tools your staff use.** Coding assistants, chat apps and agents.
- **The cost of failed requests.** See what errors are costing you.
- **Recommendations to spend less.** Where a cheaper model or a cache would do the same job, drawn from your own usage.

## Budgets and limits

- **Budgets at every level.** For the company, teams, people and projects.
- **Hard stop or warning.** Your choice, budget by budget.
- **Ask for more budget.** Staff request it, a budget owner approves.
- **Move budget between teams.** When priorities change, without new contracts.
- **Spending caps per key.** For each app, agent or person.
- **Daily limits per person.** Stops one person or agent running away with the bill.
- **Request rate limits.** Protects capacity everyone shares.
- **Alerts where you work.** By email or into your own systems.

## Who can use which AI

- **Approved models only.** Allow or block models company-wide.
- **Keys for people, projects and apps.** Each with its own rules.
- **Single sign-on.** With Microsoft Entra, Okta or Google.
- **Roles.** Owners, key admins, budget delegates and members.
- **In step with HR.** Leavers lose access automatically. People on leave are paused.
- **Work accounts only.** Your company domain, verified.

## Your data stays home

- **Prompts and answers never reach us.** Including documents and tool results. They go only to the AI companies you choose.
- **Your AI accounts' keys stay in your cloud.** We never hold them.
- **Nothing of ours in the path.** Your gateway keeps working if we're offline.
- **Support only when you allow it.** Time-limited and read-only.
- **Your staff list stays in your cloud.** We see an ID per person. Your team still sees names.
- **Delete your data.** If you close your account, your data's encryption keys are destroyed. Billing records are kept as tax law requires.

## Freedom to choose your AI company

- **Every major AI company.** OpenAI, Anthropic, Google Gemini and Vertex AI, AWS Bedrock, Azure OpenAI, and any other provider or model with a compatible API, including your own.
- **Switch without changing your apps.** Apps ask for a model by name. You decide who answers.
- **Automatic retries.** When an AI company has a hiccup.
- **A backup in an outage.** Requests move to another AI company. Staff don't notice.
- **Share traffic between AI companies.** By the weights you set.
- **Pay once for repeated questions.** Identical requests are answered from a cache in your cloud.
- **Up-to-date prices.** Including your negotiated discounts.
- **Every kind of AI request.** Text, images, audio, embeddings and more.

## AI agents and the tools they use

- **One door to company tools.** Agents reach your tools and data through the gateway (MCP), not through keys scattered across teams.
- **Tool permissions per agent.** Decide which tools each agent, team or project may use.
- **Every tool call on record.** Which agent called which tool, when, and how it went.
- **Budgets for agents too.** Spending caps and limits, as for people.
- **Shared company memory.** Assistants and agents share what your company has taught them, kept in your cloud.
- **A library of agent skills.** Versioned and checked, and given only to the teams and projects you choose.

## Safety rules

- **Keep secrets and personal data in.** Stop them being sent to AI companies.
- **Content moderation.** On what goes in and what comes out.
- **The safety services you already pay for.** AWS Bedrock Guardrails, Azure Content Safety, Google Model Armor.
- **Your own checks.** Add your company's own rules alongside ours.

## Proof for your auditors

- **Full audit log.** Filter it and export it.
- **Security questionnaires answered.** CAIQ and SIG Lite, in our security pack.
- **A data map you can review.** Every field that leaves your cloud, in our security pack.
- **Usage records in your monitoring tools.** Sent to the monitoring and log tools you already run. Never prompts or answers.

## We run it for you

- **Kubernetes, AWS ECS or one virtual machine.** Whatever your team already runs.
- **In your cloud or on your own servers.** AWS, Google Cloud, Azure, your data centre or any VPS provider.
- **Set up in one command.** Once your team approves access.
- **We monitor it.** The health of every installation, in one place.
- **We upgrade it.** Automatically, with no downtime for your staff.
- **New models and APIs, kept current.** As AI companies ship them, with no work on your side.
- **Your own image registry.** If your security team requires it.

## Try the rules yourself

Pick a request and see what your gateway does with it. (Illustrative example)

- **Support assistant:** Allowed. Approved project, within its budget.
- **Hackathon prototype:** Paused at the limit. This experiment has used its whole monthly budget, so requests wait until next month or until someone approves more. Warn only: Allowed, with a warning.
- **Personal chatbot key:** Blocked. This key isn't approved, so nothing was sent to any AI company.
- **Contract review agent:** Allowed. Approved project, within its budget. The agent's tool call was checked against its permissions.
- **Your main AI company has an outage:** Your main AI company is down, so the request went to your backup. Nobody noticed.

## From doubling every month to flat

What a budget per team does to AI spend that doubles every month. Once every request goes through Intramura, spend stops climbing and stays flat, while every team keeps using AI.

- **2× a month** spend growth before
- **Flat** once budgets are in place
- **Every team** keeps access to AI

Illustrative line chart of monthly AI spend over 12 months. Spend doubles every month from 3 to 48 thousand dollars by month 5, when Intramura is rolled out. From then on it stays flat at around 48 thousand dollars a month. A dashed line shows where spend would have gone had it kept doubling: 6.1 million dollars a month by month 12. Illustrative figures, not a customer's.

## You know what you spend on AI. Do you know what it buys?

Without that view, the only lever is switching AI off for whole departments. With it, you keep what pays, cap what doesn't and move budget between them.

1. **Today: One big number, no detail.** AI spend arrives as a few provider invoices and API keys scattered across teams. You know the total. You don't know who spent it, or what it bought.
2. **Step 1: track: Every request gets an owner.** All AI traffic goes through one gateway in your cloud. Each request is tagged with its team, project and cost, across every provider.
3. **Step 2: compare: Spend sits next to value.** With costs per project, you can see what pays back and what doesn't. Some projects deserve more. Some were never approved.
4. **Step 3: control: Budget moves to what works.** Budgets go up where the return is, experiments get a ceiling, and unapproved use is blocked before it reaches the invoice.

## What each person on your team gets

- **Finance.** Spend by team and project, budgets with a hard stop, and a month-end forecast. [Spend and budgets](https://intramura.com/features#f-spend)
- **Security and data protection.** Prompts, documents and keys never reach us. An audit log, and support access only when you allow it. [How we protect your data](https://intramura.com/security)
- **AI leads.** Every major AI model for every team, freedom to switch AI company, and control of the tools your agents use. [Models and agents](https://intramura.com/features#f-choice)
- **Platform team.** We deploy, monitor and upgrade it on Kubernetes, AWS ECS or a single virtual machine in your account. [Who does what](https://intramura.com/how-it-works)

## Book a demo: Find out where your AI budget goes

Tell us how your teams use AI today and which cloud you run on. We will show you what spend control looks like in your own account.

- Demo request form: https://intramura.com/features#demo
- Email: hello@intramura.com

- [Features](https://intramura.com/features)
- [Security](https://intramura.com/security)
- [How it works](https://intramura.com/how-it-works)
- [Pricing](https://intramura.com/pricing)

---

Intramura. From the Latin *intra muros*, within the walls. © 2026 Intramura. Intramura is an independent company. All trademarks belong to their owners.
