> For the complete documentation index, see [llms.txt](https://docs.otari.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.otari.ai/guides/tools/workspace-tools.md).

# Configure a workspace's tools

Each workspace decides which built-in tools its requests can use and which MCP servers they can reach. These steps are in the dashboard; the settings they change are described in [Built-in tools](https://docs.otari.ai/reference/gateway/tools) and [MCP](https://docs.otari.ai/reference/gateway/mcp).

A workspace setting can only narrow what the deployment allows: it cannot allow web access the deployment blocks, or turn on code execution when the deployment has no sandbox.

## Configure web search

1. Select the workspace in the top-left corner.
2. Go to **Tools**, then **Web search**.
3. Turn **Allow in** on or off. Off refuses both web search and web fetch for the workspace's requests, and direct searches too.
4. Under **Search with** and **Fetch with**, choose which of the organization's keys the workspace uses, or the organization's default, or the deployment's default tool. Only a workspace owner or admin can choose; any other member sees the list read-only.
5. Under **Who runs provider web search** and **Who runs provider web fetch**, choose whether a provider's own search and fetch tools run on the provider or on Otari, or leave them **Not set** for the request and the deployment to decide.
6. Under **Advanced**, set **Max results for this workspace** to cap the results each search returns, use **Allowed domains** to limit searches and fetched pages to some sites, or **Blocked domains** to keep them away from some, and set the options of the provider the workspace's searches run on.

The policy these steps set is described in [Per-workspace web-access policy](https://docs.otari.ai/reference/gateway/tools#per-workspace-web-access-policy), the keys in [Organization search keys](https://docs.otari.ai/reference/gateway/tools#organization-search-keys), and who runs a provider's own tools in [Web-search interception](https://docs.otari.ai/reference/gateway/tools#web-search-interception). The purpose hint is set through the API.

## Configure code execution

1. Select the workspace in the top-left corner.
2. Go to **Tools**, then **Code execution**.
3. Turn **Allow in** on or off.
4. Under **Who runs provider code tools**, choose whether a provider's own code tool or Otari's sandbox runs the code.

The other fields of the policy, such as the iteration cap, the execution timeout and the sandbox image, are set through the API.

The policy these steps set is described in [Per-workspace code policy](https://docs.otari.ai/reference/gateway/tools#per-workspace-code-policy), and who runs the code in [Code-execution executor](https://docs.otari.ai/reference/gateway/tools#code-execution-executor).

## Add an MCP server

1. Select the workspace in the top-left corner.
2. Go to **Tools**, then **MCP servers**, and click **Add MCP server**.
3. Enter a **Name**, unique in the workspace. Its tools are named after it.
4. Enter the server's **URL**, which the gateway must be able to reach, and an **Authorization token** if the server needs one.
5. Optionally, add a **Purpose hint** saying when the model should use the server, for example "Use for repository and issue lookups."
6. Under **Allowed tools**, list the tools the model may call, separated by commas, or leave it blank to allow every tool the server offers.
7. Leave **Enabled** on, and click **Add MCP server**.

Requests in the workspace can use the server's allowed tools from then on. Turning **Enabled** off keeps the server's settings but stops requests from using it. How a request reaches these servers, and their limits, are in [Workspace-scoped servers](https://docs.otari.ai/reference/gateway/mcp#workspace-scoped-servers).


---

# 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.otari.ai/guides/tools/workspace-tools.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.