Network & Mocking
Inspect network traffic, mock API responses, and test offline behavior. Inspecting requests is part of core and always available. Mocking and network state control require the network capability.
{
"mcpServers": {
"playwright": {
"command": "npx",
"args": ["@playwright/mcp@latest", "--caps=network"]
}
}
}
Inspect network requests
browser_network_requests
Returns a numbered list of network requests since loading the page. Part of core — no extra capability needed.
| Parameter | Type | Required | Description |
|---|---|---|---|
static | boolean | no | Include successful static resources like images, fonts and scripts. Defaults to false |
filter | string | no | Only return requests whose URL matches this regexp, e.g. /api/.*user |
filename | string | no | Save the list to a file instead of returning it as text |
→ browser_network_requests { filter: "api" }
1. [GET] https://api.example.com/me => [200] OK
2. [POST] https://api.example.com/users/create => [201] Created
3. [GET] https://api.example.com/settings => [200] OK
Note: 14 static requests not shown, run with "static" option to see them.
browser_network_request
Returns full details — general info, headers and bodies — of a single request, using the number
printed by browser_network_requests.
| Parameter | Type | Required | Description |
|---|---|---|---|
index | number | yes | 1-based index of the request |
part | string | no | Return only this part: request-headers, request-body, response-headers, response-body |
filename | string | no | Save the output to a file instead of returning it as text |
→ browser_network_request { index: 2 }
#2 [POST] https://api.example.com/users/create
General
status: [201] Created
duration: 45ms
type: fetch
mimeType: application/json
Request headers
content-type: application/json
...
Response headers
content-type: application/json
...
Call browser_network_request with part="request-body" to read the request body.
Call browser_network_request with part="response-body" to read the response body.
→ browser_network_request { index: 2, part: "response-body" }
{"id":42,"name":"Alice"}
Binary response bodies are written to the output directory and the tool returns the file path.
Mock API responses
browser_route
Set up a route to mock network requests matching a URL pattern. When status or body is given the
request is fulfilled with that response; otherwise the request continues with the header
modifications applied.
| Parameter | Type | Required | Description |
|---|---|---|---|
pattern | string | yes | URL pattern to match, e.g. **/api/users, **/*.{png,jpg} |
status | number | no | HTTP status code to return (default 200) |
body | string | no | Response body (text or JSON string) |
contentType | string | no | Content-Type header, e.g. application/json |
headers | string[] | no | Headers to add to the request, in "Name: Value" format |
removeHeaders | string | no | Comma-separated list of header names to remove from the request |
Mock an API endpoint
You: Mock the /api/users endpoint to return two test users.
→ browser_route {
pattern: "**/api/users",
status: 200,
body: "[{\"id\":1,\"name\":\"Alice\"},{\"id\":2,\"name\":\"Bob\"}]",
contentType: "application/json"
}
→ browser_navigate { url: "https://app.example.com/users" }
- heading "Users" [level=1] [ref=e2]
- list [ref=e4]:
- listitem [ref=e5]: Alice
- listitem [ref=e6]: Bob
Test error handling
You: Test what happens when the API returns a 503 error.
→ browser_route { pattern: "**/api/users", status: 503 }
→ browser_navigate { url: "https://app.example.com/users" }
- heading "Something went wrong" [level=1] [ref=e2]
- button "Retry" [ref=e5]
→ browser_take_screenshot
Remove the mock and verify recovery:
→ browser_unroute
→ browser_click { target: "e5" }
- heading "Users" [level=1] [ref=e2]
Block resources
→ browser_route { pattern: "**/*.jpg", status: 404 }
→ browser_route { pattern: "**/analytics/**", status: 204 }
Add or strip request headers
→ browser_route { pattern: "**/api/**", headers: ["X-Debug: 1"] }
→ browser_route { pattern: "**/api/**", removeHeaders: "cookie,authorization" }
Conditional mocking with code
For complex scenarios — delays, conditional responses, request body inspection — use browser_run_code_unsafe:
→ browser_run_code_unsafe {
code: "async (page) => {
await page.route('**/api/search', async route => {
const url = new URL(route.request().url());
const query = url.searchParams.get('q');
await route.fulfill({
body: JSON.stringify(query === 'empty' ? [] : [{ title: 'Result: ' + query }])
});
});
}"
}
Manage routes
browser_route_list
Lists all active routes. Takes no parameters.
→ browser_route_list
1. **/api/users (status=200, body=[{"id":1,"name":"Alice"}]..., contentType=application/json)
2. **/*.jpg (status=404)
3. **/analytics/** (status=204)
browser_unroute
| Parameter | Type | Required | Description |
|---|---|---|---|
pattern | string | no | URL pattern to unroute. Omit to remove all routes |
Test offline mode
browser_network_state_set
Sets the browser network state. When offline, all network requests fail.
| Parameter | Type | Required | Description |
|---|---|---|---|
state | string | yes | online or offline |
→ browser_network_state_set { state: "offline" }
→ browser_navigate { url: "https://app.example.com" }
- heading "No internet connection" [level=1] [ref=e2]
→ browser_network_state_set { state: "online" }
Restricting origins
Independently of mocking, the server can be told which origins the browser may reach at all:
["@playwright/mcp@latest", "--allowed-origins=https://example.com;http://localhost:*"]
["@playwright/mcp@latest", "--blocked-origins=https://ads.example.com"]
Both take semicolon-separated lists, and the blocklist is evaluated first. These are convenience guardrails, not a security boundary — they do not affect redirects.