> For the complete documentation index, see [llms.txt](https://docs.revenium.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/revenium-sidekick.md).

# Revenium Sidekick

Revenium Sidekick helps everyone who uses Claude Code or Claude Cowork get more from your AI budget. Inside every session it shows people what their work costs, points out costly habits as they happen

Revenium Sidekick helps everyone at your company who uses Claude Code or Claude Cowork get more from your AI budget. No one wants to waste money. They just cannot see the cost from where they work. Sidekick puts the cost there, inside every session, and does three jobs: it shows each person what their own work costs, it points out costly habits as they happen, and it holds the budgets and ticket rules you set in Revenium.

This page is for the administrator or engineering leader who decides whether to roll Sidekick out. Revenium Sidekick is included in every paid Revenium plan. To get it, ask your Revenium customer success manager, who also provides the installation guides. If you are evaluating Revenium, [talk to sales](https://docs.revenium.io/plans-and-contact/talk-to-sales.md).

## Inform: each person sees their own cost

Every developer sees their own budgets and costs as they work, with no dashboard to open and no one to ask.

* **A status line under every prompt.** In a Claude Code terminal, a line under the prompt shows how much of each of the developer's own budgets is used and when it resets, the conversation size, the prompt cache, the reasoning effort when it is above the model's default, and what the session has cost so far. A budget turns yellow as it nears its limit.
* **A morning card on their own habits.** At the first session of the day, a card sets yesterday's conversation sizes against the developer's own 30-day average.

<figure><img src="https://136027078-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FXLC45kYkVPv2AtO7QLvg%2Fuploads%2Fgit-blob-e86d35674fb7c70ccf8c7b91c6481bf901c474ba%2Frevenium-sidekick-status-line.png?alt=media" alt="An illustration of a Claude Code terminal with the Revenium status line under the prompt. Early in the day it reads: day (resets 6 PM) $62 of $300, week $540 of $1,500, ctx 82K, cache warm, session $6.40, all in green or plain. Later in the day the daily budget reads $255 of $300 and the conversation size ctx 239K, both in yellow."><figcaption><p>An illustration of the status line early in the day and later in the day. Yellow marks a budget near its limit and a conversation past 200K tokens.</p></figcaption></figure>

## Advise: costly habits pointed out as they happen

Some of the cost of AI work is hidden in how a chat is used. Sidekick points it out at the moment it matters, in plain words. Users get warnings and short notes that help them avoid common mistakes like resuming sessions when the cache has expired, running large context windows, and using model effort levels above the vendor's default.

## Control: budgets and work tickets

Sidekick holds the rules you set in Revenium inside the session, before the money is spent.

* **Budgets warn first, then stop.** A [budget](https://docs.revenium.io/track-and-control-costs/cost-controls.md) can apply to each person, to a department or to certain models. As spend nears the budget's early warning line, the person sees a warning and keeps working. When a budget set to block is used up, Claude Code refuses the next prompt or tool call with a message that says why, when the budget resets and what to do. A budget stops only the work whose spend it counts, so a budget for another tool, agent or key does not stop a developer.
* **Start in shadow mode.** A new budget can start in shadow mode, where it does not actively block, and Revenium records when it would have stopped work. Watch a pilot group for a week, then switch the budget to live.
* **Your own wording.** Each message names the tool, says what happened and gives one thing to do. You can show your own tool name in place of Revenium's, and write the message on each budget, such as who to ask for more.
* **Every session tied to its ticket.** Sidekick's ticket gate reads the ticket from the branch name or the prompt, or Claude asks the developer, and the cost of the work that follows is tied to that ticket. You choose whether the gate only reminds developers or holds their prompts until a ticket is linked.

## What budget owners see in Revenium

The same figures each person sees roll up in Revenium for the people who own the budget.

* **Spend by developer, team and model** on the [Claude Code Dashboard](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-code/claude-code.md), with budgets set once in [Guardrails](https://docs.revenium.io/track-and-control-costs/cost-controls.md) and applied in every session.
* **What each ticket cost** on [Tickets](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/tickets.md), and what each pull request cost on [PR Health](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/pr-health.md).
* **Each developer named on their usage, from one company-wide deployment.** Supported on Amazon Bedrock, Google Vertex, Microsoft Foundry or an Anthropic API key.
* **One settings page for what people see.** In Revenium, under **Settings**, the **Sidekick** page sets what developers see in the morning card and on their status line.

## Where it works

* **Claude Code** in a terminal and in the Claude desktop app's Code tab.
* **Claude Cowork tasks that run on a Mac.**
* **Claude Code cloud sessions**, which run in Anthropic's cloud.
* **Claude Cowork tasks in Anthropic's cloud.** Budgets apply once an administrator links your Claude organization to your Revenium team on the Providers page.
* **macOS and Linux.** Windows 10 and 11 are supported in beta.
* **Claude Code 2.1.295 or later.**

## How you deploy it

Sidekick builds on Revenium's tracking of your Claude Code and Claude Cowork usage, so set that up first, as [Setup Claude Code](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-code.md) and [Setup Claude Cowork](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-cowork.md) explain.

## Related

* [Guardrails](https://docs.revenium.io/track-and-control-costs/cost-controls.md): create the budgets Sidekick applies.
* [Tickets](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/tickets.md): what each work ticket cost.
* [Setup Claude Code](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-code.md) and [Setup Claude Cowork](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-cowork.md): send usage to Revenium.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend/revenium-sidekick.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
