Skip to content

armgit / cua

Public Imported from GitHub
0

Scale computer-use 2.0 with open-source drivers, cross-OS sandboxes, and benchmarks for training, evaluation, and data generation.

docs: add model and harness guides (#3069)

Francesco Bonaccie4080723395a
diff --git a/docs/content/docs/how-to-guides/driver/connect-your-agent.mdx b/docs/content/docs/how-to-guides/driver/connect-your-agent.mdx
index 4d4204d7be..069719e404 100644
--- a/docs/content/docs/how-to-guides/driver/connect-your-agent.mdx
+++ b/docs/content/docs/how-to-guides/driver/connect-your-agent.mdx
@@ -8,7 +8,8 @@ import { Callout } from 'fumadocs-ui/components/callout';
Cua Driver lets a [computer-use agent](/concepts/what-is-computer-use) drive the **host desktop**: installed apps, signed-in browser sessions, local files opened in apps, and the current OS user session. MCP-capable agents connect through `cua-driver mcp`; skill-based agents such as Prime Agent call the CLI directly.
<Callout type="info">
- This connects an agent to Cua Driver on the current machine. To create a disposable local desktop, use [Cua Sandbox](/tutorials/your-first-local-sandbox) instead.
+ This connects an agent to Cua Driver on the current machine. To create a disposable local desktop,
+ use [Cua Sandbox](/tutorials/your-first-local-sandbox) instead.
</Callout>
## Before you start
@@ -51,10 +52,10 @@ directly, and it applies only to a runtime this command launches. An agent canno
widen any of this from a tool call.
<Callout type="warn">
- A `standard`-mode runtime allows input against every application on the
- desktop. If an agent should reach only a reviewed set of apps, origins, and
- directories, register it against a `bounded` runtime instead — see [Write a
- bounded manifest](/how-to-guides/driver/write-a-bounded-manifest).
+ A `standard`-mode runtime allows input against every application on the desktop. If an agent
+ should reach only a reviewed set of apps, origins, and directories, register it against a
+ `bounded` runtime instead — see [Write a bounded
+ manifest](/how-to-guides/driver/write-a-bounded-manifest).
</Callout>
## Generate the client config
@@ -168,16 +169,16 @@ Restart Cursor and confirm `cua-driver` appears in the MCP server list.
`mcp-config` also prints the right shape for clients that use a config file or a different add command.
-| Client | Generate with | Notes |
-|---|---|---|
-| Antigravity | `cua-driver mcp-config --client antigravity` | Paste into `~/.gemini/config/mcp_config.json`; `--client gemini` is a legacy alias. |
-| OpenClaw | `cua-driver mcp-config --client openclaw` | Normal gateway-spawned MCP does not inherit OpenClaw.app's macOS permission grants; embedded hosts should use [Embedding](/reference/cua-driver/embedding). |
-| OpenCode | `cua-driver mcp-config --client opencode` | Configure a real MCP server so screenshots are preserved in image blocks. |
-| Hermes | `cua-driver mcp-config --client hermes` | Paste under `mcp_servers` and reload MCP servers in Hermes. |
-| Pi | `cua-driver mcp-config --client pi` | Pi does not support MCP natively; use one-shot `cua-driver call …` commands from its shell. |
-| Qwen Code | `cua-driver mcp-config --client qwen` | Supports both a CLI add command and `~/.qwen/settings.json`. |
-| Factory Droid | `cua-driver mcp-config --client droid` | Supports CLI and JSON config forms. |
-| ZCode | `cua-driver mcp-config --client zcode` | Configure MCP in the GUI, or use `zai mcp add` for Z.ai's separate CLI. |
+| Client | Generate with | Notes |
+| ------------- | -------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| Antigravity | `cua-driver mcp-config --client antigravity` | Paste into `~/.gemini/config/mcp_config.json`; `--client gemini` is a legacy alias. |
+| OpenClaw | `cua-driver mcp-config --client openclaw` | Normal gateway-spawned MCP does not inherit OpenClaw.app's macOS permission grants; embedded hosts should use [Embedding](/reference/cua-driver/embedding). |
+| OpenCode | `cua-driver mcp-config --client opencode` | Configure a real MCP server so screenshots are preserved in image blocks. |
+| Hermes | `cua-driver mcp-config --client hermes` | Paste under `mcp_servers` and reload MCP servers in Hermes. |
+| Pi | `cua-driver mcp-config --client pi` | Pi does not support MCP natively; use one-shot `cua-driver call …` commands from its shell. |
+| Qwen Code | `cua-driver mcp-config --client qwen` | Supports both a CLI add command and `~/.qwen/settings.json`. |
+| Factory Droid | `cua-driver mcp-config --client droid` | Supports CLI and JSON config forms. |
+| ZCode | `cua-driver mcp-config --client zcode` | Configure MCP in the GUI, or use `zai mcp add` for Z.ai's separate CLI. |
## Generic MCP JSON
@@ -202,6 +203,29 @@ It returns:
After saving the config, restart the client and confirm the `cua-driver` server is connected.
+## Clients without a dedicated preset
+
+Some MCP clients can run Cua Driver directly even though `mcp-config` does not yet have a named
+preset for them. Resolve the absolute executable path first with `command -v cua-driver`, then use
+the client's native command:
+
+```bash
+# Grok Build
+grok mcp add cua-driver -- /absolute/path/to/cua-driver mcp
+grok mcp doctor cua-driver
+
+# Kimi Code CLI
+kimi mcp add cua-driver -- /absolute/path/to/cua-driver mcp
+kimi mcp test cua-driver
+```
+
+These commands use each client's documented stdio MCP interface and the standard `cua-driver mcp`
+entry point. See the focused guides for [Grok Build](/use-cua-with/grok-build) and [Kimi
+Code](/use-cua-with/kimi-code).
+
+T3 Code is a control plane for other coding agents rather than a separate MCP client. Configure Cua
+Driver in the underlying harness, then launch that harness from [T3 Code](/use-cua-with/t3-code).
+
## Next steps
- [Keep Cua Driver running](/how-to-guides/driver/keep-running): keep the daemon alive across reboots and sessions.
diff --git a/docs/content/docs/index.mdx b/docs/content/docs/index.mdx
index 73757f0fea..fdfaa57427 100644
--- a/docs/content/docs/index.mdx
+++ b/docs/content/docs/index.mdx
@@ -49,10 +49,12 @@ These recordings show agents operating desktop apps while the user's active wind
## How these docs are organised
-Tutorials: learn by doing. Start with [Tutorials](/tutorials).
+Get Started: learn by doing. Start with [Get Started](/tutorials).
Concepts: understand how and why Cua works. Start with [Concepts](/concepts).
+Use Cua with: choose a model provider, agent harness, or local runtime. Start with [Use Cua with](/use-cua-with).
+
How-to guides: recipes for a goal. Start with [How-to guides](/how-to-guides).
Reference: precise technical description. Start with [Reference](/reference).
diff --git a/docs/content/docs/meta.json b/docs/content/docs/meta.json
index 7e852b9f51..aac0eb6054 100644
--- a/docs/content/docs/meta.json
+++ b/docs/content/docs/meta.json
@@ -6,6 +6,8 @@
"---Learn---",
"tutorials",
"concepts",
+ "---Connect---",
+ "use-cua-with",
"---Do---",
"how-to-guides",
"reference"
diff --git a/docs/content/docs/use-cua-with/anthropic.mdx b/docs/content/docs/use-cua-with/anthropic.mdx
new file mode 100644
index 0000000000..44e423d5b5
--- /dev/null
+++ b/docs/content/docs/use-cua-with/anthropic.mdx
@@ -0,0 +1,26 @@
+---
+title: Anthropic / Claude
+description: Use Claude models with Cua Driver through an agent harness.
+icon: anthropic
+---
+
+Anthropic provides the Claude model family. To give Claude access to a desktop, run it in an agent harness that can call Cua Driver. The most direct path is [Claude Code](/use-cua-with/claude-code), which supports local MCP servers.
+
+| Start here when… | Recommended path |
+| -------------------------------------- | ---------------------------------------------------------------------------------------------- |
+| You use Claude Code | [Connect Claude Code to Cua Driver](/use-cua-with/claude-code) |
+| You use another MCP client with Claude | [Use the generic MCP configuration](/how-to-guides/driver/connect-your-agent#generic-mcp-json) |
+| You want an isolated desktop | [Start a local sandbox](/tutorials/your-first-local-sandbox) |
+
+Claude Code can also act as the harness for a model served locally. That is a harness feature, not a direct connection between Cua Driver and Anthropic's API.
+
+See Anthropic's [MCP documentation](https://docs.anthropic.com/en/docs/mcp) and [Claude Code CLI reference](https://docs.anthropic.com/en/docs/claude-code/cli-usage) for the upstream client behavior.
+
+<VideoDemo
+ src="https://github.com/user-attachments/assets/4159f98d-9623-40d0-a85f-37c913b1403b"
+ poster="https://github.com/user-attachments/assets/54536146-328a-431c-95b4-14a799248026"
+ title="Claude Code builds and checks a Windows app"
+ sourceUrl="https://x.com/trycua/status/2059688966085853245"
+>
+ Claude Code uses Cua Driver to operate a WPF app without taking over the user's active window.
+</VideoDemo>
diff --git a/docs/content/docs/use-cua-with/antigravity-cli.mdx b/docs/content/docs/use-cua-with/antigravity-cli.mdx
new file mode 100644
index 0000000000..cb67d6a8af
--- /dev/null
+++ b/docs/content/docs/use-cua-with/antigravity-cli.mdx
@@ -0,0 +1,36 @@
+---
+title: Antigravity CLI
+description: Connect Google's Antigravity CLI to Cua Driver over MCP.
+icon: Rocket
+---
+
+Antigravity CLI supports local stdio MCP servers. Generate the current Cua Driver configuration:
+
+```bash
+cua-driver mcp-config --client antigravity
+```
+
+Merge the generated `cua-driver` entry into the top-level `mcpServers` object in the global configuration:
+
+```text
+~/.gemini/config/mcp_config.json
+```
+
+For project scope, use `.agents/mcp_config.json` in the workspace. The configuration has this shape:
+
+```json
+{
+ "mcpServers": {
+ "cua-driver": {
+ "command": "/absolute/path/to/cua-driver",
+ "args": ["mcp"]
+ }
+ }
+}
+```
+
+Restart `agy` after changing the file, then open `/mcp` to inspect the server status. Cua Driver's legacy `--client gemini` alias prints the same Antigravity configuration.
+
+Continue with [Connect your agent to Cua Driver](/how-to-guides/driver/connect-your-agent#other-supported-clients) for runtime and permission guidance.
+
+Upstream: [Antigravity MCP documentation](https://antigravity.google/docs/mcp) and Google's [Gemini CLI migration guide](https://antigravity.google/docs/cli/gcli-migration).
diff --git a/docs/content/docs/use-cua-with/claude-code.mdx b/docs/content/docs/use-cua-with/claude-code.mdx
new file mode 100644
index 0000000000..d6f6aeb855
--- /dev/null
+++ b/docs/content/docs/use-cua-with/claude-code.mdx
@@ -0,0 +1,33 @@
+---
+title: Claude Code
+description: Connect Claude Code to Cua Driver over MCP.
+icon: claude-code
+---
+
+Claude Code can launch Cua Driver as a local stdio MCP server. Install Cua Driver first, then register the server:
+
+```bash
+claude mcp add --transport stdio cua-driver -- cua-driver mcp
+claude mcp list
+```
+
+For an absolute-path command generated by the installed Cua Driver version, run:
+
+```bash
+cua-driver mcp-config --client claude
+```
+
+Restart Claude Code or open a fresh session after adding the server. On macOS, the Cua Driver app identity that owns the runtime needs Accessibility and Screen Recording permission.
+
+Continue with [Connect your agent to Cua Driver](/how-to-guides/driver/connect-your-agent#claude-code) for permission modes, the computer-use compatibility profile, and local-model setup.
+
+<VideoDemo
+ src="https://github.com/user-attachments/assets/4159f98d-9623-40d0-a85f-37c913b1403b"
+ poster="https://github.com/user-attachments/assets/54536146-328a-431c-95b4-14a799248026"
+ title="Claude Code checks its work in a Windows app"
+ sourceUrl="https://x.com/trycua/status/2059688966085853245"
+>
+ The agent builds a WPF app, operates it through Cua Driver, patches the code, and checks it again.
+</VideoDemo>
+
+Upstream: [Claude Code MCP documentation](https://docs.anthropic.com/en/docs/mcp).
diff --git a/docs/content/docs/use-cua-with/codex.mdx b/docs/content/docs/use-cua-with/codex.mdx
new file mode 100644
index 0000000000..5cc74c5ff5
--- /dev/null
+++ b/docs/content/docs/use-cua-with/codex.mdx
@@ -0,0 +1,29 @@
+---
+title: OpenAI Codex
+description: Connect Codex to Cua Driver over MCP.
+icon: openai
+---
+
+Codex supports local MCP servers in the CLI, IDE extension, and desktop app. Generate the command with the installed Cua Driver version:
+
+```bash
+cua-driver mcp-config --client codex
+```
+
+The generated command uses the absolute executable path. Its shape is:
+
+```bash
+codex mcp add cua-driver -- /absolute/path/to/cua-driver mcp
+codex mcp list
+```
+
+Open a fresh Codex session after registration. You can also install the Cua Driver agent skill for action-selection guidance:
+
+```bash
+cua-driver skills install
+cua-driver skills status
+```
+
+Continue with [Connect your agent to Cua Driver](/how-to-guides/driver/connect-your-agent#codex) and [Install the Cua Driver agent skill](/how-to-guides/driver/install-agent-skill).
+
+Upstream: [Codex MCP documentation](https://learn.chatgpt.com/docs/extend/mcp?surface=cli).
diff --git a/docs/content/docs/use-cua-with/google.mdx b/docs/content/docs/use-cua-with/google.mdx
new file mode 100644
index 0000000000..df479c199f
--- /dev/null
+++ b/docs/content/docs/use-cua-with/google.mdx
@@ -0,0 +1,17 @@
+---
+title: Google / Gemini
+description: Use Gemini models with Cua Driver through Antigravity CLI.
+icon: gemini
+---
+
+Antigravity CLI can run local stdio MCP servers, so it can use Cua Driver as its computer-use layer.
+
+| Start here when… | Recommended path |
+| --------------------------------- | ---------------------------------------------------------------------------------------------- |
+| You use Antigravity CLI | [Connect Antigravity CLI to Cua Driver](/use-cua-with/antigravity-cli) |
+| You are migrating from Gemini CLI | Follow Google's [migration guide](https://antigravity.google/docs/cli/gcli-migration) |
+| You configure MCP with JSON | [Use the generic MCP configuration](/how-to-guides/driver/connect-your-agent#generic-mcp-json) |
+
+Google ended Gemini CLI service for individual Google AI Pro, Ultra, and free-tier accounts on June 18, 2026. Enterprise licenses and paid API-key access remain supported, but Cua recommends Antigravity CLI for new terminal setups. Cua Driver keeps `--client gemini` as a legacy alias for the Antigravity configuration.
+
+See Google's [transition announcement](https://github.com/google-gemini/gemini-cli/discussions/28017) and [Antigravity MCP documentation](https://antigravity.google/docs/mcp) for the current client behavior.
diff --git a/docs/content/docs/use-cua-with/grok-build.mdx b/docs/content/docs/use-cua-with/grok-build.mdx
new file mode 100644
index 0000000000..e7c1fec6d6
--- /dev/null
+++ b/docs/content/docs/use-cua-with/grok-build.mdx
@@ -0,0 +1,22 @@
+---
+title: Grok Build
+description: Connect xAI's Grok Build coding agent to Cua Driver over MCP.
+icon: TerminalSquare
+---
+
+Grok Build supports local stdio MCP servers. Find the installed Cua Driver path, then register it:
+
+```bash
+command -v cua-driver
+grok mcp add cua-driver -- /absolute/path/to/cua-driver mcp
+grok mcp list
+grok mcp doctor cua-driver
+```
+
+The `--` separates Grok Build's options from the Cua Driver server command. Grok Build stores user configuration in `~/.grok/config.toml`; add `--scope project` if the configuration belongs to one repository.
+
+Cua Driver currently has no dedicated Grok Build preset, so this page follows Grok Build's native MCP command and the standard `cua-driver mcp` entry point.
+
+Continue with [Connect your agent to Cua Driver](/how-to-guides/driver/connect-your-agent) for installation, permissions, and runtime modes.
+
+Upstream: [Grok Build MCP server documentation](https://docs.x.ai/build/features/mcp-servers).
diff --git a/docs/content/docs/use-cua-with/hermes.mdx b/docs/content/docs/use-cua-with/hermes.mdx
new file mode 100644
index 0000000000..14cad54d8d
--- /dev/null
+++ b/docs/content/docs/use-cua-with/hermes.mdx
@@ -0,0 +1,33 @@
+---
+title: Hermes
+description: Connect Nous Research's Hermes Agent to Cua Driver over MCP.
+icon: Waypoints
+---
+
+Hermes Agent can launch local MCP servers from `~/.hermes/config.yaml`. Generate the current configuration from Cua Driver:
+
+```bash
+cua-driver mcp-config --client hermes
+```
+
+Paste the generated block under `mcp_servers`, then run `/reload-mcp` inside Hermes. The configuration shape is:
+
+```yaml
+mcp_servers:
+ cua-driver:
+ command: '/absolute/path/to/cua-driver'
+ args: ['mcp']
+```
+
+Continue with [Connect your agent to Cua Driver](/how-to-guides/driver/connect-your-agent#other-supported-clients) for runtime and permission guidance.
+
+<VideoDemo
+ src="https://github.com/user-attachments/assets/dcc5a81f-6379-4378-a225-adaf5dfa5dcd"
+ poster="https://github.com/user-attachments/assets/15091959-0f93-4076-9338-18ef3a737091"
+ title="Four Hermes sessions share a Windows desktop"
+ sourceUrl="https://x.com/francedot/status/2061706670208897279"
+>
+ Independent Hermes agents operate separate apps through the same Cua Driver runtime.
+</VideoDemo>
+
+Upstream: [Hermes Agent documentation](https://hermes-agent.nousresearch.com/docs/).
diff --git a/docs/content/docs/use-cua-with/index.mdx b/docs/content/docs/use-cua-with/index.mdx
new file mode 100644
index 0000000000..4633230a01
--- /dev/null
+++ b/docs/content/docs/use-cua-with/index.mdx
@@ -0,0 +1,77 @@
+---
+title: Use Cua with
+description: Choose a model provider, agent harness, or local runtime for Cua Driver.
+---
+
+import { Card, Cards } from 'fumadocs-ui/components/card';
+import { SiAnthropic, SiGooglegemini, SiMeta, SiOllama, SiOpenai } from 'react-icons/si';
+import {
+ Bot,
+ BrainCircuit,
+ Cpu,
+ MoonStar,
+ Orbit,
+ PanelsTopLeft,
+ Rocket,
+ TerminalSquare,
+ Waypoints,
+} from 'lucide-react';
+
+Cua Driver is the computer-use layer. An agent harness connects to it over MCP or through the Cua Driver CLI, then gives a model screenshots, accessibility state, and input tools. Pick the part of the stack you are starting from.
+
+## Models and providers
+
+These pages show the verified harnesses and local runtimes that can put each model family in front of Cua Driver. Cua Driver does not connect directly to a model API.
+
+<Cards>
+ <Card icon={<SiAnthropic />} title="Anthropic / Claude" href="/use-cua-with/anthropic">
+ Use Claude through Claude Code or another MCP-capable harness.
+ </Card>
+ <Card icon={<SiOpenai />} title="OpenAI" href="/use-cua-with/openai">
+ Connect Codex to Cua Driver through its native MCP support.
+ </Card>
+ <Card icon={<SiGooglegemini />} title="Google / Gemini" href="/use-cua-with/google">
+ Use Antigravity CLI as the MCP-capable harness.
+ </Card>
+ <Card icon={<Orbit />} title="xAI / Grok" href="/use-cua-with/xai">
+ Connect Grok Build to Cua Driver as a local MCP server.
+ </Card>
+ <Card icon={<SiMeta />} title="Meta" href="/use-cua-with/meta">
+ Run Muse Glimmer locally with a harness such as Claude Code.
+ </Card>
+ <Card icon={<BrainCircuit />} title="Qwen" href="/use-cua-with/qwen">
+ Use Qwen Code or serve a Qwen model through a local runtime.
+ </Card>
+ <Card icon={<MoonStar />} title="Kimi" href="/use-cua-with/kimi">
+ Connect Kimi Code CLI through its MCP support.
+ </Card>
+</Cards>
+
+## Agent harnesses
+
+Harnesses own the agent loop and invoke Cua Driver. Some are tied to a model provider; others let you choose the model separately.
+
+<Cards>
+ <Card icon={<SiAnthropic />} title="Claude Code" href="/use-cua-with/claude-code" />
+ <Card icon={<SiOpenai />} title="OpenAI Codex" href="/use-cua-with/codex" />
+ <Card icon={<Rocket />} title="Antigravity CLI" href="/use-cua-with/antigravity-cli" />
+ <Card icon={<TerminalSquare />} title="Grok Build" href="/use-cua-with/grok-build" />
+ <Card icon={<Waypoints />} title="Hermes" href="/use-cua-with/hermes" />
+ <Card icon={<Bot />} title="OpenClaw" href="/use-cua-with/openclaw" />
+ <Card icon={<PanelsTopLeft />} title="T3 Code" href="/use-cua-with/t3-code" />
+ <Card icon={<BrainCircuit />} title="Qwen Code" href="/use-cua-with/qwen-code" />
+ <Card icon={<MoonStar />} title="Kimi Code" href="/use-cua-with/kimi-code" />
+</Cards>
+
+## Local runtimes
+
+<Cards>
+ <Card icon={<SiOllama />} title="Ollama" href="/use-cua-with/ollama">
+ Launch a compatible harness against a locally served model.
+ </Card>
+ <Card icon={<Cpu />} title="llama.cpp and Unsloth" href="/use-cua-with/llama-cpp">
+ Use the tested Muse Glimmer GGUF path on Apple Silicon.
+ </Card>
+</Cards>
+
+If you already know which agent you want to use, go straight to [Connect your agent to Cua Driver](/how-to-guides/driver/connect-your-agent).
diff --git a/docs/content/docs/use-cua-with/kimi-code.mdx b/docs/content/docs/use-cua-with/kimi-code.mdx
new file mode 100644
index 0000000000..d14ddb2848
--- /dev/null
+++ b/docs/content/docs/use-cua-with/kimi-code.mdx
@@ -0,0 +1,21 @@
+---
+title: Kimi Code
+description: Connect Kimi Code CLI to Cua Driver over MCP.
+icon: MoonStar
+---
+
+Kimi Code CLI supports local stdio MCP servers. Find the installed Cua Driver path, then register and test it:
+
+```bash
+command -v cua-driver
+kimi mcp add cua-driver -- /absolute/path/to/cua-driver mcp
+kimi mcp test cua-driver
+```
+
+The `--` marks the start of the stdio command. Kimi Code can also load MCP configuration from `~/.kimi-code/mcp.json` or a project-level `.kimi-code/mcp.json`.
+
+Cua Driver currently has no dedicated Kimi Code preset, so this page follows Kimi Code's native MCP command and the standard `cua-driver mcp` entry point.
+
+Continue with the [generic MCP configuration](/how-to-guides/driver/connect-your-agent#generic-mcp-json) for Cua Driver runtime and permission guidance.
+
+Upstream: [Kimi Code CLI MCP guide](https://www.kimi.com/code/docs/en/kimi-code-cli/customization/mcp.html) and [`kimi mcp` reference](https://moonshotai.github.io/kimi-cli/en/reference/kimi-mcp.html).
diff --git a/docs/content/docs/use-cua-with/kimi.mdx b/docs/content/docs/use-cua-with/kimi.mdx
new file mode 100644
index 0000000000..d469ed6a8f
--- /dev/null
+++ b/docs/content/docs/use-cua-with/kimi.mdx
@@ -0,0 +1,15 @@
+---
+title: Kimi
+description: Use Kimi models with Cua Driver through Kimi Code CLI.
+icon: MoonStar
+---
+
+Kimi Code CLI supports local stdio MCP servers. Configure Cua Driver in the CLI, then let Kimi Code invoke its desktop tools as part of the agent loop.
+
+| Start here when… | Recommended path |
+| ----------------------------------- | -------------------------------------------------------------------------------------------------------------------------------- |
+| You use Kimi Code CLI | [Connect Kimi Code to Cua Driver](/use-cua-with/kimi-code) |
+| You manage MCP with a project file | Add Cua Driver to `.kimi-code/mcp.json` using the [generic MCP shape](/how-to-guides/driver/connect-your-agent#generic-mcp-json) |
+| You are building another Kimi agent | Use the [MCP tools](/reference/cua-driver/mcp-tools) or Cua Driver CLI |
+
+See Kimi Code CLI's [MCP customization guide](https://www.kimi.com/code/docs/en/kimi-code-cli/customization/mcp.html) and [`kimi mcp` reference](https://moonshotai.github.io/kimi-cli/en/reference/kimi-mcp.html).
diff --git a/docs/content/docs/use-cua-with/llama-cpp.mdx b/docs/content/docs/use-cua-with/llama-cpp.mdx
new file mode 100644
index 0000000000..355776b215
--- /dev/null
+++ b/docs/content/docs/use-cua-with/llama-cpp.mdx
@@ -0,0 +1,19 @@
+---
+title: llama.cpp and Unsloth
+description: Serve a local GGUF model with llama.cpp and connect it to Cua Driver.
+icon: Cpu
+---
+
+`llama.cpp` gives you direct control over local model serving. Unsloth publishes quantized GGUF variants that can reduce the memory required to run supported models.
+
+The recorded Muse Glimmer demos use:
+
+```text
+Unsloth UD-Q4_K_XL GGUF → llama.cpp → Claude Code → filtered Cua Driver MCP → macOS Lume guest
+```
+
+Follow [Run a local model with Cua Driver](/how-to-guides/driver/run-with-local-model#path-2-use-llamacpp-with-an-unsloth-gguf) for the tested `llama-server` flags and harness configuration.
+
+The local-model guide also covers the main optimization: reduce the MCP tool catalog before the model sees it. Tool schemas, screenshots, and accessibility trees all consume context. Keep state reads bounded and expose only the actions required by the task.
+
+Upstream: [`llama.cpp`](https://github.com/ggml-org/llama.cpp), its [server documentation](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md), [function-calling guide](https://github.com/ggml-org/llama.cpp/blob/master/docs/function-calling.md), and Unsloth's [Muse Glimmer guide](https://unsloth.ai/docs/models/muse-glimmer).
diff --git a/docs/content/docs/use-cua-with/meta.json b/docs/content/docs/use-cua-with/meta.json
new file mode 100644
index 0000000000..05df730496
--- /dev/null
+++ b/docs/content/docs/use-cua-with/meta.json
@@ -0,0 +1,28 @@
+{
+ "title": "Use Cua with",
+ "icon": "Cable",
+ "pages": [
+ "index",
+ "---Models & providers---",
+ "anthropic",
+ "openai",
+ "google",
+ "xai",
+ "meta",
+ "qwen",
+ "kimi",
+ "---Agent harnesses---",
+ "claude-code",
+ "codex",
+ "antigravity-cli",
+ "grok-build",
+ "hermes",
+ "openclaw",
+ "t3-code",
+ "qwen-code",
+ "kimi-code",
+ "---Local runtimes---",
+ "ollama",
+ "llama-cpp"
+ ]
+}
diff --git a/docs/content/docs/use-cua-with/meta.mdx b/docs/content/docs/use-cua-with/meta.mdx
new file mode 100644
index 0000000000..64da94fce4
--- /dev/null
+++ b/docs/content/docs/use-cua-with/meta.mdx
@@ -0,0 +1,36 @@
+---
+title: Meta
+description: Run Meta computer-use models locally with Cua Driver.
+icon: meta
+---
+
+Meta's Muse Glimmer 30B is an open-weight vision-language model designed for local agentic computer use. A runtime serves the model, an agent harness owns the loop, and Cua Driver supplies screenshots, accessibility state, and input actions.
+
+The tested Cua path uses:
+
+- Meta's Muse Glimmer 30B model
+- Unsloth's `UD-Q4_K_XL` GGUF
+- `llama.cpp` on Apple Silicon
+- Claude Code as the harness
+- Cua Driver running inside a macOS Lume guest
+
+Follow [Run a local model with Cua Driver](/how-to-guides/driver/run-with-local-model) for the complete setup, including the smaller tool surface used to control context growth.
+
+See Meta's [Muse Glimmer model card](https://huggingface.co/meta-models/Muse-Glimmer-30B) and Unsloth's [Muse Glimmer guide](https://unsloth.ai/docs/models/muse-glimmer) for the upstream model and quantization details.
+
+<div className="not-prose my-8 grid gap-5 md:grid-cols-2">
+ <VideoDemo
+ src="https://cuademo06261324.blob.core.windows.net/artifacts/demos/muse-glimmer/notes-editorial.mp4"
+ poster="https://cuademo06261324.blob.core.windows.net/artifacts/demos/muse-glimmer/notes-editorial-poster.jpg"
+ title="Create a checklist in Apple Notes"
+ >
+ Muse Glimmer runs locally and completes the task through Cua Driver.
+ </VideoDemo>
+ <VideoDemo
+ src="https://cuademo06261324.blob.core.windows.net/artifacts/demos/muse-glimmer/reminders-editorial.mp4"
+ poster="https://cuademo06261324.blob.core.windows.net/artifacts/demos/muse-glimmer/reminders-editorial-poster.jpg"
+ title="Schedule an Apple Reminder"
+ >
+ Muse Glimmer sets the reminder and verifies the final state.
+ </VideoDemo>
+</div>
diff --git a/docs/content/docs/use-cua-with/ollama.mdx b/docs/content/docs/use-cua-with/ollama.mdx
new file mode 100644
index 0000000000..22c5243a59
--- /dev/null
+++ b/docs/content/docs/use-cua-with/ollama.mdx
@@ -0,0 +1,19 @@
+---
+title: Ollama
+description: Serve a local model with Ollama and connect it to Cua Driver through an agent harness.
+icon: ollama
+---
+
+Ollama serves local models. It does not replace the agent harness or connect to Cua Driver by itself. Use an Ollama-supported harness for the agent loop, then register Cua Driver in that harness.
+
+For Muse Glimmer on Apple Silicon, the tested stack is:
+
+```text
+Muse Glimmer in Ollama → Claude Code → filtered Cua Driver MCP → desktop
+```
+
+Follow [Run a local model with Cua Driver](/how-to-guides/driver/run-with-local-model#path-1-use-ollama-on-apple-silicon) for the exact model, context, and Claude Code launch command.
+
+Local computer use consumes context quickly. Start with a model that supports images and tool calling, expose only the Cua Driver tools the task needs, and use a context window of at least 64K. The tested guide uses 128K.
+
+Upstream: [Ollama's Claude Code integration](https://docs.ollama.com/integrations/claude-code) and [tool-calling documentation](https://docs.ollama.com/capabilities/tool-calling).
diff --git a/docs/content/docs/use-cua-with/openai.mdx b/docs/content/docs/use-cua-with/openai.mdx
new file mode 100644
index 0000000000..204ff3aebe
--- /dev/null
+++ b/docs/content/docs/use-cua-with/openai.mdx
@@ -0,0 +1,17 @@
+---
+title: OpenAI
+description: Use OpenAI models with Cua Driver through Codex.
+icon: openai
+---
+
+OpenAI Codex supports local MCP servers in its CLI, IDE extension, and desktop app. Register Cua Driver once, then use its tools from a Codex session.
+
+| Start here when… | Recommended path |
+| ------------------------------------------ | ---------------------------------------------------------------------------------------------- |
+| You use Codex | [Connect Codex to Cua Driver](/use-cua-with/codex) |
+| You use another MCP-capable OpenAI harness | [Use the generic MCP configuration](/how-to-guides/driver/connect-your-agent#generic-mcp-json) |
+| You are building your own agent | [Read the MCP tool reference](/reference/cua-driver/mcp-tools) |
+
+Cua Driver supplies desktop state and actions. Codex owns the model interaction, reasoning loop, and tool selection.
+
+See OpenAI's current [Codex MCP documentation](https://learn.chatgpt.com/docs/extend/mcp?surface=cli) for supported transports and configuration.
diff --git a/docs/content/docs/use-cua-with/openclaw.mdx b/docs/content/docs/use-cua-with/openclaw.mdx
new file mode 100644
index 0000000000..1d6a39b588
--- /dev/null
+++ b/docs/content/docs/use-cua-with/openclaw.mdx
@@ -0,0 +1,29 @@
+---
+title: OpenClaw
+description: Connect OpenClaw to Cua Driver over MCP.
+icon: Bot
+---
+
+import { Callout } from 'fumadocs-ui/components/callout';
+
+OpenClaw can register Cua Driver as a local MCP server. Generate the current command:
+
+```bash
+cua-driver mcp-config --client openclaw
+```
+
+The generated command uses the absolute Cua Driver executable path:
+
+```bash
+openclaw mcp set cua-driver '{"command":"/absolute/path/to/cua-driver","args":["mcp"]}'
+```
+
+<Callout type="warn">
+ On macOS, a gateway-spawned MCP process does not inherit OpenClaw.app's Accessibility and Screen
+ Recording grants. Grant permissions to the identity that actually runs Cua Driver, or use the
+ documented embedding path.
+</Callout>
+
+Continue with [Connect your agent to Cua Driver](/how-to-guides/driver/connect-your-agent#other-supported-clients). To run OpenClaw inside a Cua-managed Linux VM, use [Run OpenClaw in a Lume VM](/how-to-guides/lume/run-openclaw).
+
+Upstream: [OpenClaw MCP CLI documentation](https://docs.openclaw.ai/cli/mcp).
diff --git a/docs/content/docs/use-cua-with/qwen-code.mdx b/docs/content/docs/use-cua-with/qwen-code.mdx
new file mode 100644
index 0000000000..f2c25bef16
--- /dev/null
+++ b/docs/content/docs/use-cua-with/qwen-code.mdx
@@ -0,0 +1,24 @@
+---
+title: Qwen Code
+description: Connect Qwen Code to Cua Driver over MCP.
+icon: BrainCircuit
+---
+
+Qwen Code supports local stdio MCP servers. Generate the current command with Cua Driver:
+
+```bash
+cua-driver mcp-config --client qwen
+```
+
+The generated command uses the absolute executable path:
+
+```bash
+qwen mcp add cua-driver /absolute/path/to/cua-driver mcp
+qwen mcp list
+```
+
+You can also add the generic `mcpServers` object to `~/.qwen/settings.json`. Restart Qwen Code after changing the configuration.
+
+Continue with [Connect your agent to Cua Driver](/how-to-guides/driver/connect-your-agent#other-supported-clients) for runtime and permission guidance.
+
+Upstream: [Qwen Code MCP documentation](https://qwenlm.github.io/qwen-code-docs/en/users/features/mcp/).
diff --git a/docs/content/docs/use-cua-with/qwen.mdx b/docs/content/docs/use-cua-with/qwen.mdx
new file mode 100644
index 0000000000..1c900cd13d
--- /dev/null
+++ b/docs/content/docs/use-cua-with/qwen.mdx
@@ -0,0 +1,17 @@
+---
+title: Qwen
+description: Use Qwen models with Cua Driver through Qwen Code or a local runtime.
+icon: BrainCircuit
+---
+
+Qwen models can reach Cua Driver through an MCP-capable harness or a local serving stack.
+
+| Start here when… | Recommended path |
+| ------------------------------ | ---------------------------------------------------------------------------------------------------------------- |
+| You use Qwen Code | [Connect Qwen Code to Cua Driver](/use-cua-with/qwen-code) |
+| You serve a Qwen model locally | Choose [Ollama](/use-cua-with/ollama) or [llama.cpp](/use-cua-with/llama-cpp), then connect a compatible harness |
+| You use another MCP client | [Use the generic MCP configuration](/how-to-guides/driver/connect-your-agent#generic-mcp-json) |
+
+The model and computer-use layer remain separate. Qwen Code or another harness selects tools; Cua Driver observes and acts on the desktop.
+
+See Qwen Code's [MCP documentation](https://qwenlm.github.io/qwen-code-docs/en/users/features/mcp/).
diff --git a/docs/content/docs/use-cua-with/t3-code.mdx b/docs/content/docs/use-cua-with/t3-code.mdx
new file mode 100644
index 0000000000..ef635beda1
--- /dev/null
+++ b/docs/content/docs/use-cua-with/t3-code.mdx
@@ -0,0 +1,19 @@
+---
+title: T3 Code
+description: Use Cua Driver from an agent harness managed in T3 Code.
+icon: PanelsTopLeft
+---
+
+T3 Code is a control plane for coding agents. It runs and organizes harnesses such as Claude Code, Codex, Qwen Code, and Grok Build; it is not itself the MCP client that launches Cua Driver.
+
+To use Cua Driver from T3 Code:
+
+1. Choose one of T3 Code's supported agent harnesses.
+2. Register Cua Driver in that harness using its page in this section.
+3. Start or restart the harness from T3 Code and confirm the Cua Driver tools are available.
+
+This separation keeps the MCP configuration portable. The same harness can use Cua Driver whether it is launched from T3 Code or directly from a terminal.
+
+Start with [Claude Code](/use-cua-with/claude-code), [OpenAI Codex](/use-cua-with/codex), [Grok Build](/use-cua-with/grok-build), or [Qwen Code](/use-cua-with/qwen-code).
+
+Upstream: [T3 Code](https://t3.codes/) and the [T3 Code repository](https://github.com/pingdotgg/t3code).
diff --git a/docs/content/docs/use-cua-with/xai.mdx b/docs/content/docs/use-cua-with/xai.mdx
new file mode 100644
index 0000000000..2bf1c7886c
--- /dev/null
+++ b/docs/content/docs/use-cua-with/xai.mdx
@@ -0,0 +1,17 @@
+---
+title: xAI / Grok
+description: Use Grok models with Cua Driver through Grok Build.
+icon: Orbit
+---
+
+Grok Build is xAI's coding agent and supports local MCP servers. It is the shortest verified route from a Grok model to Cua Driver.
+
+| Start here when… | Recommended path |
+| ----------------------------------------- | -------------------------------------------------------------------------------- |
+| You use Grok Build | [Connect Grok Build to Cua Driver](/use-cua-with/grok-build) |
+| You are building a custom Grok agent | Call Cua Driver through [MCP](/reference/cua-driver/mcp-tools) or the CLI |
+| You want to manage the agent from T3 Code | Configure Grok Build first, then [use it through T3 Code](/use-cua-with/t3-code) |
+
+Cua Driver does not call the Grok API directly. Grok Build owns the agent loop and launches Cua Driver as a local tool server.
+
+See xAI's [Grok Build overview](https://docs.x.ai/build/overview) and [MCP server guide](https://docs.x.ai/build/features/mcp-servers).
diff --git a/docs/src/lib/source.ts b/docs/src/lib/source.ts
index d4a3d58fd4..73d99e97a4 100644
--- a/docs/src/lib/source.ts
+++ b/docs/src/lib/source.ts
@@ -1,13 +1,19 @@
import { docs } from '@/.source';
import { loader } from 'fumadocs-core/source';
import { icons } from 'lucide-react';
-import { createElement } from 'react';
+import { cloneElement, createElement } from 'react';
+import { SiAnthropic, SiGooglegemini, SiMeta, SiOllama, SiOpenai } from 'react-icons/si';
import fs from 'node:fs/promises';
import path from 'node:path';
// Custom brand icons (not available in Lucide)
const brandIcons: Record<string, () => React.ReactElement> = {
+ anthropic: () => createElement(SiAnthropic, { className: 'w-4 h-4' }),
+ gemini: () => createElement(SiGooglegemini, { className: 'w-4 h-4' }),
+ meta: () => createElement(SiMeta, { className: 'w-4 h-4' }),
+ ollama: () => createElement(SiOllama, { className: 'w-4 h-4' }),
+ openai: () => createElement(SiOpenai, { className: 'w-4 h-4' }),
github: () =>
createElement(
'svg',
@@ -110,9 +116,9 @@ const brandIcons: Record<string, () => React.ReactElement> = {
function iconResolver(icon: string | undefined) {
if (!icon) return;
// Check brand icons first (github, discord, etc.)
- if (icon in brandIcons) return brandIcons[icon]();
+ if (icon in brandIcons) return cloneElement(brandIcons[icon](), { key: icon });
// Fall back to Lucide icons
- if (icon in icons) return createElement(icons[icon as keyof typeof icons]);
+ if (icon in icons) return createElement(icons[icon as keyof typeof icons], { key: icon });
}
/**