Built for: Anyone who needs to believe we can build the hardest-looking software there is.
The problem
Trading software is the stress test of interface engineering: hundreds of numbers changing every second, charts that must redraw without stutter, and money math that has to be exact while everything moves. Most portfolios show a screenshot of one. This one runs.
What we built
A candlestick chart drawn from scratch. Canvas, ninety bars, volume underneath, a crosshair that reads out OHLC as your finger moves, three timeframes that regenerate the series.
An order book that never sleeps. Ten levels each side with depth bars, a live spread, and a time-and-sales tape printing while you watch.
Orders with consequences. Route a market order, watch it work, get the fill, and see the position appear with profit and loss recomputing on every tick.
Decisions we made
Prices come from a random walk with volatility clustering, the same math real simulators use, so the chart moves like a market rather than static.
Invented tickers by design. Real ones would imply real data we do not have.
On Monday morning
This is the reference app: if the studio can keep sixty numbers coherent at 800 millisecond ticks on a phone, the dashboard your business actually needs is a quiet afternoon.
What to try in the demo
Hover the chart and flip 1M to 1D
Open BOOK and watch the tape print
Buy 100 AXOM, wait for the fill, then find it in positions repricing live
[email protected]Tell us what you are building. We answer within one business day.
Built for: Any business that rolls vans: plumbers, couriers, field engineers, deliveries.
The problem
Dispatch by phone is a memory game. Who is closest, who is finishing, who is already late. The answer lives in one dispatcher’s head, and when they are off, it lives nowhere.
What we built
A city that is alive. Six vehicles patrolling a vector map, each one tappable for its card: driver, status, ETA.
Dispatch with evidence. A burst main pulses red until someone claims it. The sheet ranks free vehicles by live ETA; pick one and watch it change course, arrive, and flip the task to on-site.
The board behind the map. Every task with its countdown, every driver with their day: on-time rate, stops, route, event log.
Decisions we made
Real map tiles need external servers, which this page forbids by policy, so the city is a designed vector. The movement, ETA and dispatch mechanics are exactly what a production build runs on GPS.
One task completes on screen mid-demo, because real ops boards are never static.
On Monday morning
The burst-pipe call gets dispatched by whoever answers the phone, with the veteran’s knowledge built into the screen: nearest crew, honest ETA, one tap.
What to try in the demo
Tap a moving vehicle for its card
Tap the red pin, compare ETAs, dispatch
Stay on the map until the van arrives, then read the task timeline
[email protected]Tell us what you are building. We answer within one business day.
Built for: Any company putting an AI agent in front of customers and wondering what it actually does all day.
The problem
An AI agent in production is a black box that spends money and talks to your customers. When it fails, you need to see exactly which step failed and why; when you improve it, you need proof the new prompt is actually better. That tooling barely exists, and we build it.
What we built
Every run, dissected. A waterfall of spans: agent, retriever, model calls, tool calls, each with its timing bar and its actual payload, expandable.
Failure made visible. One trace shows a rate-limited tool call in red, and the retry that saved it. Error handling is the part that proves production experience.
Evals, not vibes. Prompt v3 against v4 across six criteria and 120 replayed conversations, with the sample answers side by side. The rollout decision becomes evidence.
Decisions we made
Costs shown at real precision and latencies log-normal, because anyone who runs agents knows what those distributions look like.
The replay streams token by token, the way the model actually answered.
On Monday morning
When your agent tells a customer something wrong, you find the exact span that caused it before lunch, and the eval table tells you whether the fix helped or just felt better.
What to try in the demo
Open the run marked ERROR and find the retry
Replay the answer stream on an llm span
Flip V3 to V4 in evals and read what actually improved
[email protected]Tell us what you are building. We answer within one business day.
Built for: Software and product teams, and any client who names Linear in the brief.
The problem
Issue trackers are where software teams live, and the good ones feel instant: keyboard-first, zero loading states, every mutation everywhere at once. Clients cite Linear by name. Matching that feel is the test.
What we built
One array of truth. Twenty-six issues drive board, list, counts and analytics; move one anywhere and everything repaints together.
The palette. Cmd K opens a command bar that filters sixteen real commands and executes them: assign, reprioritize, move to Done, jump views. With a toast to prove it happened.
A cycle that reads honest. Burndown with ideal against actual, a scope-change annotation, and velocity compared to last cycle.
Decisions we made
Issue titles are real engineering sentences, comment counts follow a power law, and timestamps cluster in working hours, because reviewers check exactly these things.
On Monday morning
Your team stops paying for four tools that do not talk to each other, and the Monday standup runs off one board that was already true before the meeting.
What to try in the demo
Tap an issue, cycle its status pill, watch the board agree
Open Cmd K, type "done", execute
Filter the list to Urgent and find the two heavy conversations
[email protected]Tell us what you are building. We answer within one business day.
Built for: Any business moving money online, and any client who names Stripe in the brief.
The problem
Payment operations live in the gap between "the chart says fine" and "this specific charge is fraud". The tool has to hold both: the aggregate view and the single suspicious payment with its evidence, plus the review queue where a human decides.
What we built
Books that reconcile. The volume tile, the chart and the payments table all sum from the same data, to the cent. Check it; reviewers do.
Risk with reasons. Every flagged payment shows the rule that caught it and the checks behind the score: CVC, address, geography mismatches. Approve or decline, and every count on every screen moves.
The boring 85%. Most payments succeed unremarkably, because real books are boring, and boring is what credible looks like.
Decisions we made
Payment IDs are shaped exactly like the real API’s, and failure reasons are the ones that exist, because fintech buyers notice.
On Monday morning
The chargeback that used to eat a morning becomes a two-minute read: the timeline, the signals, the rule it tripped, and a decision button.
What to try in the demo
Filter payments to Needs review
Open one and copy its ID, then read its risk checks
Clear the whole review queue and meet the empty state
[email protected]Tell us what you are building. We answer within one business day.
Built for: Any product or marketing team drowning in events and starving for answers.
The problem
Dashboards answer yesterday’s questions. Analytics tools earn their keep when you can ask a new question and get a chart back in seconds: rebuild the funnel, split by plan, read the retention triangle. That interactivity is the product.
What we built
A funnel you can edit. Remove a step, add another, and the conversion math recomputes and re-animates, always decreasing downstream, because funnels that leak upward are lies.
The retention triangle. Eight cohorts decaying convexly to their plateau, every cell tappable for its exact number.
Segmentation that splits live. One line becomes three by plan, and the table under it matches the chart exactly.
Decisions we made
Series carry weekend dips and gentle growth because flat generated data is the fastest tell of a fake dashboard.
A 300 millisecond shimmer on re-query, because instant reads as fake in analytics.
On Monday morning
The "should we build annual billing" argument ends in four taps: funnel to upgrade, split by plan, read the drop-off, decide with numbers.
What to try in the demo
Remove a funnel step and watch the math recompute
Tap cells in the retention triangle
Split weekly actives by plan, then flip the dashboard to 90 days
[email protected]Tell us what you are building. We answer within one business day.
Built for: Law firms and any regulated practice where the software must know the rules.
The problem
Legal software is a domain-knowledge test. Trust money must never mix with operating money, deadlines are court-fatal, time is billed in tenths, and estate intakes hide the classic failure: a beneficiary form that quietly outranks the will.
What we built
The whole practice on one screen. Matters with unbilled work, deadlines counting down in red, and trust funds in a dashed box that never sums with anything else. That separation is the law, drawn.
Time that becomes money. A running timer files entries in tenths at each person’s rate; the billable toggle moves totals instantly; one tap drafts an invoice that equals the time tab exactly.
The intake module. The estate questionnaire runs the conflict engine on the client’s own answers: add a retirement account with no beneficiary and watch it get caught, then file the flags to the matter.
Decisions we made
The intake collects and organizes; the attorney does the legal work. No document generation, no advice, by design.
On Monday morning
The intake that took three weeks of chasing happens on the client’s couch, arrives pre-flagged, and every minute spent on it is already on the invoice draft.
What to try in the demo
Start the timer, stop it, find the entry it filed
Open the estate matter, run the intake, add an IRA with no beneficiary
Generate the invoice and check it equals the time tab
[email protected]Tell us what you are building. We answer within one business day.
Built for: Any brand selling physical product, and any client who names Shopify in the brief.
The problem
Commerce operations is two truths per order (paid? shipped?) times a thousand orders, plus stock that lies unless the tool keeps it honest. Back offices win on density done calmly.
What we built
Orders that read at a glance. Payment and fulfillment as independent statuses on every row, because that duality is real domain modeling, filter tabs with live counts, and search across everything.
Bulk work made one gesture. Check four orders and fulfill them together; pills, tabs and counts all move at once.
Inventory you can touch. Tap a stock number, step it, watch the low-stock warning appear or clear.
Decisions we made
One coherent 12-product catalog with variants and believable prices, one refund and one payment-pending order, because clean books read fake.
On Monday morning
The morning fulfillment pile becomes one filtered view and one bulk action, and stock stops surprising you on Thursdays.
What to try in the demo
Filter to Unfulfilled, select several, fulfill them together
Open an order and add an internal note to its timeline
Edit a stock level inline and trip the low-stock warning
[email protected]Tell us what you are building. We answer within one business day.
Built for: Teams whose knowledge lives in eleven places, and any client who names Notion in the brief.
The problem
Editors are the interaction class engineers fear most, which is exactly why one belongs in this portfolio. Real block editing, real formatting, and a database that shows the same records as a table or a board.
What we built
A real editor, honestly sourced. Built on Quill, the open-source rich-text engine, credited in this page’s source as its license requires. The one component you never hand-build; knowing that is part of the craft.
Slash commands and a toolbar that formats. Type / on an empty line for headings, lists, quotes, code. Select text and bold it. It simply works.
AI where it earns its place. Ask AI reads the open page and streams a summary block into the document, highlighted, yours to edit or delete.
Decisions we made
Three seeded pages of believable company writing, onboarding, roadmap, meeting notes, because an empty editor demos nothing.
On Monday morning
The answer to "where is that decision written down" becomes one place, and the new person reads the onboarding page instead of interrupting four people.
What to try in the demo
Type / on an empty line and insert a heading
Ask AI for the page summary and watch it stream in
Open Launch tasks and flip table to board
[email protected]Tell us what you are building. We answer within one business day.
Built for: Teams that think on walls, and any client who names FigJam or Miro in the brief.
The problem
The multiplayer canvas is the most memeable interaction on the modern web: pan, zoom, drag, and other people’s cursors moving with names attached. Building one in a browser, smoothly, on a phone, is the flex.
What we built
A canvas that feels infinite. Pan by dragging, pinch or button zoom with correct math at every scale, a Fit button that frames the whole board.
Stickies that behave. Add, drag, edit, recolor, delete, with a connector that re-anchors live as its sticky moves.
Colleagues, staged honestly. Anna and Dimitris drift, pause, drag a sticky, and type a new one character by character. Scripted theater, labeled as such, and the mechanics are the real thing.
Decisions we made
The seeded board is a believable sprint retro, because an empty canvas demos nothing and lorem stickies insult the viewer.
On Monday morning
The workshop that needed a conference-room wall happens with the remote half of the team actually present, on the same board, with names on their cursors.
What to try in the demo
Pan and zoom, then hit Fit
Drag a sticky and watch its connector follow
Add your own sticky and recolor it, then watch Dimitris add his
[email protected]Tell us what you are building. We answer within one business day.
Vantage MarketsLIVE · --:--:--
AXOMAxiom Industrial0.00+0.00%
Order BookAXOM
SizePriceSize
SPREAD 0.02
Time & sales
Order TicketAXOM
100
MARKET
—
WORKING · BUY 100 AXOM · ROUTED
SymQtyAvgLastuP&L
Open P&L0.00
Watchlist8 SYMBOLS
Meridian Fleet--:--
VAN-081 · Dimitris
Tap a vehicle for its card; tap the red pin to dispatch. Invented city, live mechanics.
Task Board
Task Detail
Task
Recipient
Driver
ETA
Driver
Driver
0%On time
0Stops today
0hOn shift
Route, next stops
Event log
Dispatch: burst main, Vine St
Free vehicles, ETA from live positions
TracelineLAST 7 DAYS
RunLatencyTokensCost
Trace
MonitorALBA SUPPORT AGENT
1.8sp50 latency
11.6sp99 latency
3.9%Error rate
Latency, 7 days
p50p99
Cost by model, this week
sonnet-4-5$41.20
haiku-4-5$17.05
gpt-4o-mini$6.14
Tokens per day
EvalsPROMPT V3 VS V4 · 120 CASES
Scores are means over 120 replayed support conversations, judged per criterion. V4 wins grounding and tone; V3 still answers faster. The rollout call is yours, with the evidence in one place.
Board
Meridian Issues · Cycle 14 ·
Issues
Activity
Cycle 14
Scope
Completed
Velocity
Ideal Actual Scope change
Enter runs the top result · acting on
LedgerlineFathomly Analytics · live
Gross volumeLast 30 days
Gross volume
–
12.4% vs last 30d
MRR
–
3.2% vs last 30d
Dispute rate
0.08%
1 of 1,240 charges, 90d
0.02pp vs last 90d
Next payout
–
–
8.1% vs last payout
Payments
Last 30 days
–
pi_
Timeline
Risk
0
Customer
Metadata
Review queue
Rule matches wait here for a decision
Queue clear
Radar keeps watching.
Copied
Northbeam · Inkwell
Funnel
Signed up to upgraded · last 30 days · all platforms
Add a step
Northbeam · Inkwell
Retention
Weekly cohorts · share still active each week
9%100% of cohort active
Tap a cell for the exact share. Week 0 is the signup week, so it always reads 100%.
Northbeam · Inkwell
Segment
Weekly active users · last 8 weeks
Northbeam · Inkwell
Dashboards
Growth review · 6 saved charts
Hartwell & Vanetimer idle
€0Unbilled WIP
€12,940Outstanding
€41,200Trust held
Trust money is client money. It lives in its own accounting and never mixes with operating funds; the dashed box is the visual law of this app.
Deadlines
Matters
Mattertimer idle
Matter
Client
Unbilled
Trust balance
Next deadline
Activity
Timetimer idle
0:00:00Vasileiou estate · drafting
Entries today
Billable today
Billingtimer idle
€4,1200-30d
€5,61031-60d
€2,41061-90d
€80090d+
Invoices
Pulls today’s billable entries for the Vasileiou estate into a draft, totalled by each person’s rate.
Estate intakeclient-facing
Anna’s family, as she gave it.
The client answered on her phone over a weekend. Confirm and move to assets.
Assets, with their beneficiary forms.
A form on an account outranks the will. Add one and watch the check react.
Checking…
Alba SupplyTODAY
0 selected
Order
Items
Subtotal
Shipping
VAT 24%
Total
Fulfillment
Timeline
Inventory
InsightsLAST 30 DAYS
Sales per day, last 30 days
Top products
Storefront
8,412Sessions
6.1%Add to cart
2.4%Checkout
Fieldnote
Favorites
Workspace
FieldnoteWorkspace / Team onboarding
Team onboarding
Edited by George, just now · type / on an empty line for blocks