Enque Changelog

Enque Connect, and audits that go deep #

Run a small desktop app and the platform reads your local crawls — no ports, no exposure — while every audit workbook gains a ranked action plan and five more audits ship.

Your crawler, reachable from the platform — and the audits that read it got a lot deeper:

• Enque Connect (preview) — a small agent you run on the machine where Screaming Frog lives opens one outbound connection to Enque, so audits running on the server can read your local crawls. No port forwarding, no exposing Screaming Frog to the internet, nothing inbound. Pair once with a code from the Tools page; it stores a token and reconnects on its own, and your workspace's connection is written for you, so you never touch a URL or key.
• It's a desktop app now — instead of a terminal command, download EnqueConnect, run it, and paste your pairing code once. It lives in the system tray (green when it's relaying, red when it's reconnecting), starts on login, and reconnects by itself. No Python, no terminal to keep open — the script path is still there for developers. The app is served from the platform's own storage, so Tools → Screaming Frog → Connect shows a Download for Windows button the moment a build is published, and never a dead link before that.
• Connect from the Tools page, and see who has — open the Screaming Frog connection, click Connect, and the panel flips to Connected on its own once the tunnel is up. Because one user runs one licence, the tool no longer shows an agency-wide key form nobody fills: each member connects their own crawler, and the agency admin console shows who is connected and how many are online.
• Big crawls read faster and never time out — reading a large crawl through Enque Connect now reaches the tunnel by an internal route instead of going back out through the public network, removing a hard 100-second ceiling that could fail a big export and shaving latency off every read.
• Audit workbooks got deep — every SEO audit report now opens with a computed headline and a ranked Action Plan: every check, worst first by severity, with the pages affected, an authored fix, the effort, the exact rule applied and example URLs. Each check gets its own drill-down tab holding the affected pages with their evidence; a clean check states "verified clean on N pages" and an unmeasured one says NOT MEASURED on its face, so an empty tab is a statement, never a blank. No model writes a single cell.
• Four more audits from the same crawl, no extra setup — Orphan Pages lists the pages the crawler knows about from the sitemap, Analytics or Search Console that nothing on the site links to; Response Time & Server Performance flags every page by how slowly the server answered it (Slow over 1s, Very slow over 3s) from the measured value, never an estimate; Redirect Chain & 3xx lists every redirect the crawl followed — chains, loops, refreshes, internal and external — with what each lands on and the internal links that should point at the destination directly; and Robots & Blocked Resources surfaces everything robots.txt blocks, with the blocked JS, CSS and image resources called out as the dangerous case. Alongside them, Site Architecture & Crawl Overview gives the shape of the crawl at a glance — URLs found and crawled, the internal/external split, and how much of the site is actually indexable, each with its share of the total. A composition summary, not a scored audit; the specific issues live in the other audits.
• Competitive Landscape & Proposal — one consolidated, client-ready proposal built from every completed marketplace and quick-commerce audit a client has: competitor archetypes, cross-channel gaps and a workstream plan, with every figure computed from the audit runs.
• Audits stop asking which site to audit — a workspace already names its site, yet every audit's first run asked again. The run configuration now resolves it from the workspace's own record and says so in the run summary, asking only when there genuinely is no answer. Competitor lists, locations and keywords are still questions, because they are.
• The audit wave holds up against real crawls — all eighteen element audits were run end to end against a live crawl and every failure fixed: a column present in one crawl and not another is skipped and named in the report instead of failing the whole audit, a site with nothing of a given type produces an honest zero-page report instead of an error, and an audit of an element carrying no indexability data no longer claims every page is non-indexable.
• Fixes — saving a connection's site scope no longer wipes the tunnel URL and token behind it, so the connection can't vanish the moment you touch the sites; a report chip's renderer re-syncs on every seed pass, so a workspace onboarded in an older version no longer fails when an audit finishes; and a run summary can no longer read its own freshly landed rows as "no data".