> ## Documentation Index
> Fetch the complete documentation index at: https://linkly.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Using Linkly AI in OpenClaw

> Integrate your Linkly AI local knowledge base with the OpenClaw open-source AI Agent platform

## What You Can Do

> [accordion] Retrieve private local documents during autonomous Agent decision-making
>
> OpenClaw's Agent will autonomously call Linkly AI's `search` tool to search your computer for relevant knowledge based on the task at hand — no need to manually specify where the materials are. The entire workflow runs truly autonomously.

> [accordion] Use three tools to go deeper layer by layer and efficiently complete complex tasks
>
> From `search` to find relevant documents, `outline` to quickly grasp the structure, to `read` for careful reading of key chapters — OpenClaw's Agent pipeline can fully leverage all three Linkly AI tools to efficiently complete research and analysis tasks.

> [accordion] Let OpenClaw handle the installation
>
> Run `clawhub install linkly-ai` to install the Skills, and the Skills will guide you through the setup.

## Prerequisites

* [OpenClaw](https://openclaw.org) installed
* Linkly AI Desktop running with the MCP service active

## Configuration Steps

> [tab] CLI (Recommended)
>
> First [install Linkly AI CLI](https://linkly.ai/docs/en/use-cli.md) in OpenClaw. Once installed, OpenClaw can directly call CLI commands to search and read your documents:
>
> ```bash theme={null}
> linkly search "keyword"
> linkly read <document-id>
> ```
>
> You can also install Skills with one command via ClawHub and let the Skills guide you through the setup — simpler and easier:
>
> ```bash theme={null}
> clawhub install linkly-ai
> ```
>
> After installation, OpenClaw will automatically configure the Linkly AI connection. You can start using `linkly search`, `linkly read`, and other commands to search and read your documents.

> [tab] MCP HTTP
>
> Manually add the MCP server in OpenClaw:
>
> **Local use:**
>
> * URL: `http://127.0.0.1:60606/mcp`
> * No additional authentication required
>
> **LAN use:**
>
> * URL: `http://192.168.x.x:60606/mcp` (use your actual LAN IP)
> * Header: `Authorization: Bearer <your access token>`
> * Requires turning on **Allow LAN access** on the **Agent usage** page of the Linkly AI Desktop main window; the address and token are in the **LAN** section of **View MCP connection settings**
>
> **Cloud use:**
>
> Requires completing [remote access configuration](https://linkly.ai/docs/en/use-in-remote.md) first. The cloud connection is authorized in the browser, with no key to fill in:
>
> ```bash theme={null}
> openclaw mcp set linkly-ai '{"url":"https://mcp.linkly.ai/mcp","transport":"streamable-http","auth":"oauth"}'
> openclaw mcp login linkly-ai
> ```
>
> `openclaw mcp login` prints an authorization link: open it in a browser, sign in to linkly.ai and click **Allow**. Check the connection with `openclaw mcp status --verbose`. To stop using it later, run `openclaw mcp logout linkly-ai`, or revoke it under **Authorized apps** on the [Security page of the dashboard](https://linkly.ai/dashboard/security).
>
> > [!NOTE]
> >
> > Enter the LAN access token into the config yourself — don't send it to an AI assistant to fill in: conversations are sent to the model provider and kept in the chat history. If it has already appeared in an AI conversation, regenerate it under **View MCP connection settings**.

> [!TIP]
>
> We recommend also installing [Linkly AI Skills](https://linkly.ai/docs/en/use-skills.md) to teach the AI assistant how to efficiently use Linkly AI's tools, improving search and reading efficiency.

## FAQ

> [accordion] clawhub install fails
>
> Make sure the ClawHub CLI is correctly installed and your network can reach openclaw\.org. You can also choose the MCP HTTP method for manual configuration.

> [accordion] Cannot connect to Linkly AI
>
> Please check: - Whether Linkly AI Desktop is running - Whether the MCP service is active - In LAN mode, confirm the access token is correct; in cloud mode, run `openclaw mcp status --verbose` to check the authorization, and run `openclaw mcp login linkly-ai` again if it has lapsed


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.