CoreDash vs Akamai mPulse

mPulse has been around longer than most RUM tools. It started at SOASTA, Akamai bought SOASTA, and the app still logs in at mpulse.soasta.com. It runs on boomerang.js, one of the first open-source RUM libraries, which is still maintained. If your site is on the Akamai CDN you might already have it, because new properties get a free Lite app by default.

I built CoreDash, so keep that in mind. I'll try to be fair, which with mPulse isn't difficult.

What Akamai mPulse does well

Start with the install. On the Akamai CDN you add the mPulse behavior in Property Manager and the snippet gets injected at the edge. No template change, no ticket for the dev team. Most RUM tools (CoreDash too) need someone to paste a script tag first. It also works outside Akamai delivery with a normal loader snippet, so you're not tied to their CDN.

akamai.com, October 2026.

The other strong point is money. You can track revenue and conversion as custom metrics next to your performance data, and mPulse then runs what-if forecasts on what a faster page would be worth. I've seen plenty of performance projects get stuck in a meeting where someone asks what the fix is worth. mPulse gives you an answer with numbers from your own shop. It also does conversion impact and A/B test comparisons, and puts a cost on your third-party resources.

The Core Web Vitals attribution is decent. For LCP you get the element selector, tag, src and srcset, the size, and since March 2024 the transfer time and decoded size of the LCP resource. INP and CLS come with the element behind them. On single page apps mPulse reports an incremental INP per soft navigation, which is harder to get right than you'd think. Boomerang also handles details people often get wrong: prerendered pages are corrected for activationStart, back/forward cache restores are handled, and there's a PCI mode that keeps it away from card fields.

It's made for big setups: more than 200 metrics, real-time dashboards, automatic anomaly detection, bot traffic kept separate from people, 13 months of data on Enterprise, a REST API and a Datadog integration. According to reviewers non-technical people can find their way around the UI, which isn't true for every enterprise tool.

If mPulse is already running

If mPulse measures your real visitors today, you're better off than most sites I look at. Those are still guessing from a Lighthouse score on the mobile tab and waiting for CrUX to tell them something broke weeks ago. Collecting data from every real visitor in every browser is the most important part, and mPulse does that. If it answers your questions, keep it. The rest of this page is about the questions it's less good at, and what that costs you after a year or two.

Side by side

Pricing and features checked September 2026.

Where CoreDash goes further

Seeing that LCP got worse is just the start. Most of the work is finding out why and getting the fix live, and that's the part CoreDash helps with.

Which element, and why it was slow

mPulse tells you which image was the LCP element and how long the download took. Useful, but in my experience the download is rarely the slow part. More often the browser found the image late (it's a CSS background, or it was lazy loaded), or the image was ready and the page just hadn't painted it yet. That's load delay and render delay, and the mPulse beacon doesn't have a field for either.

CoreDash splits every LCP into its four phases. It also records the LCP priority, so you can see right away if the hero image was preloaded, had a high fetchpriority or was lazy loaded. The Diagnosis dashboard shows which phase is too slow and what the fix looks like on your stack.

Same for INP. mPulse points at the element that was slow to respond, but doesn't split that time into input delay, processing and presentation, and it doesn't use Long Animation Frames (the boomerang Continuity plugin works with Long Tasks). A Long Task only tells you the main thread was busy. It has no field for the script. On most sites I look at that's the tag manager, a chat widget or a consent script. CoreDash stores the LoAF script URL with every slow interaction, plus the input type and load state, so you can group INP by script and see which vendor hurts the most. My Long Animation Frames article explains how.

Your AI agent can read the data

Most of my own debugging starts with an agent reading the field data now, and for that the agent needs to be able to query it. Akamai publishes an llms.txt for the mPulse docs, which is more than most vendors do, but that only covers the docs. There's no official mPulse MCP server. Akamai's own MCP servers are for Akamai Cloud and Akamai Functions. Someone in the community built one on top of the Query API, so it can be done, but then you rely on a side project and on having the API, which mPulse Lite doesn't include.

The CoreDash MCP server is included in every plan, Starter too. It has three tools: get_metrics for the p75 (or p50 up to p95) with the good/needs improvement/poor split per dimension, get_timeseries to see if a metric is improving or regressing, and get_histogram for the distribution. Claude connects as a custom connector with OAuth (setup guide), and Claude Code, Cursor, Windsurf, VS Code and Gemini CLI work too. CWV Superpowers is a free MIT-licensed agent that combines that field data with a Chrome trace and gives you a diff with the file, the line and the fix. The same endpoint is also the plain API, so getting Core Web Vitals into your own reports is just POSTing JSON and getting JSON back.

Prices on the website

mPulse doesn't publish prices. You talk to sales, agree on a beacon commitment, and the trial goes through sales too. For a big company that already has an Akamai contract that's normal, and reviewers on PeerSpot call the pricing reasonable. For a small team or an agency it means a sales process before you've seen any of your own data. The free Lite app only works on Akamai CDN properties and keeps 30 days, which isn't enough to spot a regression that slowly builds up over a quarter.

Akamai mPulse docs on Lite apps, October 2026.

CoreDash has its prices on the pricing page, self-serve checkout all the way up to Enterprise and a 14-day trial without a credit card. You pay for pageviews, and every plan has every dashboard and no limit on URLs. Retention goes from a year on Standard to eight years on Enterprise. Agencies get agency plans with pageviews pooled across clients, and the Agency API creates a project for each new client from your own tooling. mPulse's tenant and app model works well for lots of sites inside one company, but it wasn't made for an agency with thirty clients.

When to pick Akamai mPulse instead

If your site is on Akamai and mPulse is already running, keep it. RUM without a code change is worth a lot, especially in organisations where getting a script tag into the template takes a quarter.

mPulse is also the better choice if your main question is about money. Revenue next to your performance data plus what-if forecasts on what a faster page would earn is what it does best. CoreDash tracks conversions too, but doesn't forecast revenue. Same if you need PCI mode on checkout pages, if your team already lives in Datadog, or if you want 200+ metrics beyond Core Web Vitals in one enterprise contract.

You can also run both. The CoreDash script is under 2KB and async, so it runs fine next to boomerang. mPulse keeps doing the business reporting and you use CoreDash to debug Core Web Vitals.

Try it next to mPulse

Add the CoreDash snippet next to mPulse and give it two weeks on your own traffic. Look at the LCP phases and LoAF scripts for your slowest templates, and connect Claude to the MCP server to see what an agent does with your data. The trial is 14 days, no credit card. Check pricing, or book a demo and I'll show you around.

See also: all comparisons, CoreDash vs Datadog and CoreDash vs New Relic.

Sources: Akamai mPulse product page, mPulse Lite limits, edge injection, what is in an mPulse beacon, boomerang 1.851 release notes, Core Web Vitals by element widgets, LCP transfer time and decoded size, mPulse API, boomerang on GitHub, Akamai Cloud MCP server, PeerSpot reviews.