interact_and_test

MCP tool from Periscope MCP by segentic-lab

Multi-step workflow with 25 actions (see below)

How to use it

Project documentation lists interact_and_test for the Periscope MCP MCP server. Add the server to your MCP client (Claude Desktop, Cursor, Windsurf and others), then check which tools your installed version makes available. Tool availability can depend on configuration and credentials. A server handshake does not verify this tool’s behavior. See the full listing for setup details.

Install Periscope MCP

$pip install -r requirements.txt
FULL PERISCOPE MCP LISTING

Other tools in Periscope MCP (39)

check_links

Comprehensive link checker (internal + external)

click_element

Click element (force=true bypasses overlays)

close_session

Close session and free resources

compare_screenshots

Pixel diff between two screenshots

copy_auth

Copy auth config + session state between projects

crawl_project

Discover all pages from base URL

create_project

Create a new testing project

delete_project

Delete project + data

fill_form

Fill form fields, optionally submit

flow

Save / run / list / delete named step sequences (reusable workflows)

get_interaction_log

Export real INP time series (per interaction) as JSON/CSV + percentile stats

get_page_elements

List matching elements with attributes

get_project

Get project details

get_report

Read a report file

get_screenshot

Get screenshot file path

get_table_data

Parse HTML table into structured JSON (headers → cell values)

get_toast_messages

Capture visible toast/notification messages

interactive_login

Open a visible window to log in by hand (2FA/SSO/CAPTCHA), then savelogin

list_projects

List all projects

list_reports

List saved test reports

list_sessions

List all active sessions

login_project

Execute login using configured auth

measure_interaction

Measure click-to-result timing

open_session

Open persistent browser session (headed=true for a visible window)

run_lighthouse

Real Google Lighthouse audit: 0-100 scores, Core Web Vitals, failed audits (needs Node.js)

save_login

Capture the manual-login session; the project then runs authenticated + headless

scroll_into_view

Scroll element into viewport without clicking

select_option

Native <select> or custom dropdown (Radix/shadcn) — auto-detects

select_page

Adopt a popup/new tab (OAuth, target=blank) as a new drivable session

session_report

HTML+PDF dossier of every tool call this run — args (redacted), verdicts, timings, screenshots

set_basic_auth

Configure HTTP Basic Auth

set_cookies

Inject session cookies

set_form_login

Configure username/password form login

set_viewport

Switch viewport size (8 device presets or custom w/h)

test_form_validation

Analyze form validation messages

test_project

Full audit: crawl + test all pages

test_responsive

Test at mobile/tablet/desktop viewports

test_url

Test a single URL (screenshot + checks)

visual_check

Named visual-regression baselines: set once, check for a hard pass/fail