building custom dashboards with the edgeful API + Claude Code
summary: turn edgeful API data into clean, shareable dashboards with Claude Code and the /dashboard skill — pull reports, mash subreports together, compare tickers side by side, and let Claude surface the patterns you can't see in the UI.
the edgeful API gives you every stat behind every report — and once that data is in your workspace, you can turn it into a dashboard that's curated exactly the way you want to see it. this is where the API gets fun: you stop reading reports one page at a time and start combining them, comparing them, and asking Claude to find the patterns for you.
this article picks up after setup. if you haven't connected the API yet, start with setting up the edgeful API with VS Code + Claude Code — it walks you through installing the tools, plugging in your key, and adding the dashboard skill. come back here once you can pull data.
watch the walkthrough
what you'll need
2 things, both covered in the setup article:
the edgeful API connected in VS Code with Claude Code — so you can pull data in plain english
the edgeful dashboard skill installed in your workspace — a small file that teaches Claude Code how to turn API responses into clean HTML dashboards. you can grab it from the API docs; the setup article covers installing it in step 7
one gotcha worth repeating: the first time you add the skill, restart Claude Code so it picks up the /dashboard command. if you type /dashboard and nothing happens, that's almost always why.
the basic flow — pull, then build
every dashboard starts the same way: pull the data, then hand it to the skill.
1. pull the reports you want. ask Claude Code for the data in plain english — name the reports (or subreports), the ticker, the session, and the lookback. for example:
pull all the IB subreports for NQ in the New York session over the last year from the edgeful API
Claude Code hits the right endpoints and saves the responses into a file in your workspace. you don't have to read the raw JSON.
2. build the dashboard. point the skill at what you just pulled:
put the NQ data you pulled into /dashboard
it generates a single-file HTML dashboard — usually in a minute or two — that you can open in your browser. that's the whole loop: pull, then /dashboard.
what shows up in a dashboard
the skill doesn't just dump numbers. for each report it pulls in, the dashboard lays out:
what the report shows — a short description of what you're looking at
how to read it — so the dashboard stands on its own
the headline — the key number or takeaway
the data, visualized — break-type distributions, where price closed relative to a range, rejection stats, and so on, each rendered in a format that fits the report
the piece that ties it all together: every report on the dashboard links back to the live report on edgeful. if you see something you don't recognize — say the IB by double break stat — click it and you land on that exact report on edgeful, where you can dig into what it means. that's a big reason to use the skill rather than building from scratch: the links come for free, and they always point to the right spot.
going deeper — past a simple repurpose
a first dashboard is usually just the same report data in a cleaner layout. that's useful, but it's the floor, not the ceiling. the real unlock is asking Claude to do things the edgeful UI doesn't lay out for you.
ask it to go further. if you're not sure what you want yet, ask open-ended:
how can we make this dashboard more detailed — think outside the box, not just repurposing the same data
or flip it around and let Claude interview you:
ask me a few questions about how I trade, then build a dashboard based on my answers
that second prompt is the one to reach for when you know the data is valuable but you don't know how you want it shaped yet.
cross-report mashups
because you've pulled multiple reports and subreports into one workspace, Claude can combine them in ways the UI keeps on separate pages. a common one: take a directional subreport and slice it by a second dimension — for example, IB rejection continuation broken out by weekday, or rejection stats grouped by the size of the IB.
these mashups are where you tend to find things you'd never have gone looking for. in André's walkthrough, the standout finding wasn't about direction at all — it was about timing: the IB rejection setup held up across most weekdays but fell apart on one of them. you don't get a read like that from any single report page; it falls out of combining two.
one note: edgeful AI can also do cross-report analysis if you'd rather stay inside the platform — the API route just lets you build a persistent, visual dashboard around it.
let Claude find the patterns
once a pile of data is in the workspace, you don't have to know what you're looking for. hand the whole thing over and let Claude hunt:
act as my financial analyst and dig through all this data — flag any explicit, high-value patterns you think I should be aware of
you'll get back a prioritized read on what stands out — which setups are strong, which are uneven, and where the real edge is hiding. treat it as a starting point: confirm anything interesting against the live report (the dashboard links make that one click) before you lean on it.
comparing tickers side by side
this is the one most people can't do easily in the UI — putting two instruments on the same report, in the same view. pull both, then ask for the comparison:
pull all the IB reports for ES in the New York session over the last year, then build a new dashboard comparing ES and NQ
the dashboard puts them next to each other — single break vs. single break, double break vs. double break, rejection stats side by side — so the differences jump out instead of forcing you to flip between two report pages and hold the numbers in your head.
the pairs worth comparing this way:
ES vs NQ — the two index futures most traders watch together
SPY vs ES and QQQ vs NQ — the ETF against its futures equivalent
MNQ vs NQ — micro vs. full-size, for anyone who trades one but references the other
add YM into an index comparison for a third point of reference
anything you can pull for one ticker, you can pull for another and stack them.
tips for getting more out of it
start simple, then iterate. build a basic dashboard first, see what's there, then ask for more depth. André built the examples in the video in about 5 minutes each — spend an hour and you're well past that.
always use the skill, not a from-scratch build. the
/dashboardskill gives you the report links back to edgeful automatically — that context is what makes a dashboard actually usable later.pull subreports, not just top-line reports. the more specific the data you pull, the more curated the dashboard gets — and the more interesting the mashups become.
verify before you trade it. a dashboard is a way to see the data, not a signal. click through to the live report and confirm any pattern holds up across timeframes before it changes how you trade.
getting a dashboard that's genuinely useful to your process takes a bit of customization and iteration — the API just hands you the raw material and removes the constraints of the UI. what you build with it is up to you.
building something interesting, or have a wish-list item for the API or the skills? reach out through support — we're building this with you.