How to use PlainSight

›Tap Try sample data to explore any industry instantly — no account needed. Free also includes ONE upload of your own file: create a free account, drop in a CSV or Excel (or paste a table), and PlainSight builds your full on-screen dashboard with the on-device rules engine, saved to your account so you can return to it. Downloading reports/datasets, exporting your analysis, and the AI features are Pro (or a one-time Report pack). Ledger-photo reading uses AI, so it's Pro.

›ASK FOR WHAT YOU CAME FOR — a box on the home page you can type a plain question into: "show me what this looks like for a gym", "where do I see my worst month?", "how do I sell a bit?". It answers from this documentation and can TAKE you there — loading that industry's sample, starting the guided tour, or opening a handbook. Free, ten questions a day, no account needed. It deliberately cannot see your data: if you ask about your own numbers it says so and points at Ask your data, which is the Pro half. Comedy has the same box under the setlist, called Ask the usher.

›SHAREABLE LINKS — Two kinds, for two situations. ?sample= opens PlainSight straight into a specific industry so they can look around: https://plainsightai.app/?sample=trading opens the trading journal, ?sample=hotel the hotel, ?sample=gym the fitness sample, ?sample=oilgas the oil & gas well, and so on (salon, barber, resto, ecommerce, legal, tax, gig, household…). ?tour= does the same AND starts the guided walkthrough on it — https://plainsightai.app/?tour=hotel loads the hotel data and then narrates what's on screen and what you'd do about it. Send ?tour= to someone who has never seen the product and ?sample= to someone who just wants the numbers; the tour version is usually the right one for a cold link. ?tour=1 on its own opens the tour at its industry picker. A third form, https://plainsightai.app/?handbook=hotel, opens the same industry with the free starting-out roadmap already expanded and scrolled to and the dashboard below it — for someone researching whether to start a business rather than analysing one they already run. The 18 guide pages under /guides/ each link to their matching walkthrough, and to the handbook where that trade has one. Both skip the cover page, both work with the alias spellings (?tour=restaurant, ?sample=fitness), and a key that no longer exists falls back to the picker rather than a blank screen. Once a sample is loaded, ⌕ Jump to… has "Copy guided-tour link" and "Copy sample link" so you never have to build one by hand — they're offered only on samples, because a link built while your own file is open would open a sample for whoever you sent it to. Add &subscribe=market or &subscribe=oilgas and the app also scrolls to that newsletter's signup card once the view has loaded — that's what the buttons on the home page use.

›ENTREPRENEURIAL HANDBOOKS — Twenty industries ship a built-in roadmap for people starting one of these businesses: 🧭 Coaching & consulting, 🎙️ Podcast / creator, 💇 Salon / barber / beauty, 🏋️ Gym / fitness studio, 🍽️ Restaurant / cafe / bar, 🏨 Hotel / lodging, 🏠 Short-term rental / Airbnb, 🚛 Towing / recovery, 🛍️ E-commerce, 🚚 Trucking / freight, 🚗 Gig economy / rideshare, 🏪 Convenience / retail store, 🛒 Grocery / market, 🦷 Dental practice, 🏥 Physical therapy practice, 🚬 Cigar lounge, 🌿 Landscaping / home services, 🖋️ Tattoo studio, 🧾 Tax prep / accounting, and ⛵ Yacht / boat charter. It appears automatically whenever your industry matches — typing "hair salon", "boutique gym", "coffee shop", "tow truck company", "dentist office", "lawn care", "bookkeeping" or "corner store" on your own data is enough, you don't need to load a sample. A "Starting a…" section then appears under the dashboard containing: what the market is actually worth with sourced figures, the stages a business like this moves through and the mistake that stalls each one, what every revenue stream is genuinely good for — what it pays and whether it scales — and a numbered list of what to do in your first month. ALL OF THAT IS FREE and needs no account: it's published research and rules of thumb, and charging for it would be charging for a library card. Pro adds exactly one thing on top — "Read this against my numbers", which compares the roadmap to your loaded data and tells you which stage you're actually in, which stream is carrying you, and what's missing. Free users get a "Preview AI-enhanced sample" button showing exactly what that produces, written on example numbers and clearly labelled as such. Some verticals deliberately have NO handbook: household budget isn't a business, trading has its own dedicated view, and oil & gas production is a capital-intensive operation rather than something you start from scratch — a roadmap there would be pretending. The matcher also excludes look-alikes on purpose — "Oilfield services / consulting" doesn't get the coaching handbook, and "Private jet / air charter" doesn't get the boat one — because a confidently wrong roadmap is worse than none. Ordering matters for the same reason: "Cigar lounge" is checked before restaurants (it contains "lounge"), physical therapy before dental (both are medical practices with completely different economics), and towing before trucking. Every handbook carries an industry-specific caution about licensing and regulation, because several of these trades are licensed and the rules are usually local.

›HANDBOOK: WHAT EACH ONE LEADS WITH — Coaching: rates rise from roughly $150/hr under a year of experience toward $310+ past ten, and the arc is about breaking the hours-for-money link. Podcast: only about 1.9% of podcasts earn any ad revenue at all, so the show is the marketing and the audience is the customer. Salon: rebooking rate and retail share (target 10–15% of revenue) decide profitability more than chair count. Gym: membership is the base and rarely the profit — personal and small-group training carry it, and monthly churn is the number to watch. Restaurant: US sales look like they're growing 4.8% but that's about 1% after inflation, so you're competing for existing traffic; prime cost and the lease decide the rest. Towing: the market has been CONTRACTING about 2% a year, so contracts and storage matter more than cash calls. E-commerce: one of the few genuinely growing markets here (+10.4%), where contribution margin and acquisition cost are the only three numbers that matter. Trucking: an estimated 70–80% of new owner-operators exit within two years and 8,000+ carriers have shut down since January 2024 — cost per mile and cash flow, not rate per load, are what separate the survivors. Convenience store: the highest-selling categories (tobacco, lottery, fuel) are the worst-margin ones — they buy foot traffic, and prepared food makes the money. Grocery: a $963B market with famously thin net margins where independents win on specialty, prepared food and service, never on price against a chain. Dental: collections rather than production is the real number, and hygiene recall predicts next year's revenue better than this year's does. Physical therapy: reimbursement per visit has trended down while costs have not, which is why cash-pay wellness and performance services keep appearing. Cigar lounge: IBISWorld forecasts the category to DECLINE over the next five years, and indoor-smoking ordinances decide the location before the lease does — memberships and lockers, not retail, are the business. Landscaping: route density (jobs per mile driven) decides profitability more than jobs won, and the off-season has to be planned in your first summer. Tattoo: a small $1.3B market where clients follow the artist rather than the shop — own your portfolio and following from day one of an apprenticeship. Tax prep: roughly two-thirds of revenue arrives in four months against twelve months of overhead, and monthly bookkeeping is the standard fix everyone eventually finds. Charter: utilisation days decide everything, because the vessel costs the same whether it sails or sits. Gig / rideshare: the apps advertise $26–31 an hour gross, but after fuel, maintenance, insurance and depreciation the real figure is usually $12–19 — so dollars per MILE, not per hour, is the number that tells you whether it's working. Hotel / lodging: the booking CHANNEL decides profit more than the room rate does — an OTA takes 15–25% where a direct booking costs 4–5%, and OTAs drive about 63% of independent hotels' bookings, so moving ten points to direct changes nothing on the revenue line and drops straight to profit. The hotel sample runs four years across TWO properties, the second opening in year three, so you can see what a ramping location legitimately looks like next to a mature one. Short-term rental / Airbnb: Airbnb's host fee went from about 3% to 15.5% in 2026 and is charged on the cleaning fee too, US occupancy averages 57.4% rather than the 75–80% people model with, and the rules are LOCAL — a city ordinance can end the business on a council vote, so verify at the address before buying. The coaching sample runs a full TEN years so the income arc is visible on the trend chart: entry-rate 1:1 work, then packages, then group cohorts and workshops breaking the hours-for-money link, then corporate and the passive layers. The podcast sample runs five years and is deliberately honest about sponsorship being small next to what an audience buys directly.

›MODE INTROS — The first time in a browser session that you open a trading journal or an oil & gas / oilfield-services sample, a short cinematic clip plays over the loading view. It's atmosphere, not a gate: there's always a Continue button, it never blocks the data, it plays at most once per session per mode, and it's skipped entirely on deep links opened inside an embedded frame.

›PREFER NOT TO READ? — Tap ⌕ Jump to… in the header (or press ⌘K / Ctrl+K on a keyboard) to open the command palette: type a few letters and jump straight to a sample, a tab, or an action (export, add data, benchmark, save). Once a file or sample is loaded, a slim next-step bar under the header offers the one obvious action for where you are — dismiss it with × if you'd rather not see it. (It no longer appears before you have data: on the landing page every action it offered was already on screen.) The 💬 Ask button in the bottom-right corner takes a plain question at any time, free, ten a day, and it cannot see your file. And 🧭 Quick tour (footer, or the palette → "tour") walks you through the whole flow, only when you ask for it — it loads a sample in your industry and actually drives the app, switching tabs and highlighting each section as it goes — tap 🔊 Listen and it reads each step aloud in a natural voice, and keeps narrating as you move through; pick any voice your device has from the dropdown. Nothing is downloaded and no audio is sent anywhere — it's your browser's own speech engine.

›LANGUAGES — tap 🌐 Language on the home screen to read PlainSight in any of 100+ languages. Type to search the full list (Español, Kreyòl, 中文, Polski — anything Google Translate supports) and pick from the dropdown. It opens the app through Google Translate, so everything on the free plan works fully; for the live-AI features, use the English original.

›PlainSight reads what's in the file and picks the right view automatically — Revenue for sales and invoices, Operations for oil & gas production, or Trading for a trade log or raw broker export. A top-bar toggle appears whenever more than one view fits.

›THE FRONT PAGE — a demo reel pages through the sample library one industry at a time (hover pauses it, ◂ ▸ browse, 'stills' turns it off and is remembered; reduced-motion gets stills by default), and the Starting-out handbook rail drifts below like a shop window — tap a card to load that industry's sample, and the handbook opens from the dashboard. Every frame and card is a working door, not a screenshot.

›It auto-detects your columns by their content, not their names — no setup or mapping needed. The detected columns are shown at the bottom of the dashboard so you can verify.

›Files, pasted data, and spreadsheets are processed entirely in your browser. The one exception is ledger photos, which are sent to AI vision to be read — see Privacy below.

›Any transaction-style export: invoices, sales, payments — one row per transaction, with a date and an amount.

›Exports from QuickBooks, Toast, Square, Clover, Stripe, Shopify, Xero, PayPal, or a plain spreadsheet all work. Bank exports work too — and a personal/household budget export lands in Revenue view with spending-aware wording and household benchmarks (the 30% housing guideline, 50/30/20 savings rate). Don't see your tool? ＋ Recommend a connector sits at the end of that chip row — it opens a suggestion already written, so you name the tool and press Send.

›BIG FILES & PARQUET — .parquet files, and CSV/JSON exports over about 25 MB, are read by a built-in analytics engine (DuckDB) instead of the standard reader. It streams the file rather than loading it into memory, so warehouse-scale exports open without freezing the tab. The engine is a one-time download the first time you need it, and the app tells you that's what's happening rather than just appearing to hang. Everything still runs on your device.

›DATA IN A WAREHOUSE? — Under the connector chips, "Data in a warehouse? Connect an export" reads exports from Snowflake, BigQuery, Redshift, Postgres and Databricks. PlainSight does NOT log into your database, for two reasons: a live connection would need a relay server holding your credentials and streaming your rows through it, which would end the promise that your data stays on your device — and browsers cannot open the raw TCP sockets that Postgres and Redshift speak anyway. Instead you export a query to storage YOU control (S3, Google Cloud Storage, Azure Blob, Cloudflare R2 or Supabase), presign the object, and paste that link. The file is then read directly inside your browser tab over HTTP range requests, so a large Parquet export only transfers what's needed. Copy-paste export SQL for each warehouse is built into the panel. The signed link is stored in this browser only, is never sent to a PlainSight server, and never appears in a shareable link — treat it like a password, because anyone holding it can read that export. The panel shows when the link expires (seven days is the S3 maximum) and warns you two days out; ↻ Refresh from source re-reads the same export whenever you want current numbers. Your bucket needs a CORS rule allowing GET and the Range header from https://plainsightai.app. Postgres is simpler still — \copy writes a CSV to your own machine and you upload it like any other file.

›Using Toast POS (common for bars and restaurants)? PlainSight recognizes Toast exports automatically. In Toast Web, go to Reports and download either the item-level export (ItemSelectionDetails.csv — best, since it carries Sales Category so you can see revenue by Liquor, Beer, Hookah, Food) or the order export (OrderDetails.csv). Drop the CSV in as-is: PlainSight maps Toast's own column names, uses the net (post-discount) revenue, groups by sales category, and automatically excludes voided items — no renaming or cleanup needed. A banner confirms what it detected.

›For trading: a journal export with a per-trade P/L column (TradeZella, Tradervue, or your own sheet) loads straight into Trading view. A raw broker export with no P/L — Robinhood, Public, ThinkOrSwim, Interactive Brokers — also works: PlainSight pairs the individual buy/sell fills into round-trip trades for you.

›Multi-sheet Excel: pick the sheet with your data, tap Combine all sheets when tabs share the same columns (e.g. one tab per month), or — for oil & gas — load a specific tab as completion data if production and completion live in the same workbook.

›COMBINED BOOK (Pro) — Analyze several files together. On a computer, 📁 Import a folder reads a whole directory of broker exports at once, subfolders included — it groups files by their column shape rather than by folder (real archives keep several brokers in one place), splits thinkorswim AccountStatement files apart to find the trade history inside them, drops cancelled and failed orders, and ignores .DS_Store and PDFs. Folder picking doesn't exist on phones or tablets, so that button only appears on a computer. ＋ Add data lets you add more date-range files to the SAME account (handy when a broker only exports 365 days at a time — upload each year into the one account) or add a NEW account/location. Overlapping ranges within an account are de-duplicated automatically; identical rows across different accounts are flagged as possible double-counting. Every row is tagged with its Source, so you can filter or break performance out by account. Files with different column names get a quick “match your columns” step (guessed for you, and remembered next time), so different brokerages line up under one schema. For trading, round-trip trades pair WITHIN each account, and a By-account strip shows net P/L, win rate, and profit factor per source. Same idea works for any vertical — several brokerages, store locations, or rental properties in one view.

›Oil & gas extras: alongside production, you can upload completion data (lateral length, frac stages, proppant, formation) and formation tops (Well, Formation, Top Depth — one row per formation per well) to unlock the completion panel and a wellbore drawn to your real surveyed depths. Both accept CSV or Excel and match to wells by name, and each has a Download template button (pre-filled with your well names) so you can copy the exact shape. Uploading and templating well, completion, and tops data is free — the AI reservoir benchmarks and the downloadable report are the Pro pieces.

›Column names don't need to match ours — detection reads the content. The detected columns are shown at the bottom of the dashboard so you can verify.

›The built-in sample library spans a wide range of businesses — agency, oilfield services (invoices to operators by service line and basin), restaurant & hotel, grocery, medical/dental, legal, tax, auto, towing, landscaping, rentals, a bar & hookah lounge, a tattoo shop (artists, custom vs. flash work, piercing), a barber shop (revenue by chair), a full salon & spa (stylists, color, facials, massage), a nail salon (technicians, gel, acrylics), a skin care clinic / med spa (providers, injectables, laser, memberships), a yacht charter (with add-ons like jet skis, snorkeling, and catering), a trucking & freight carrier (multiple trucks across lanes), a private jet charter (fleet with empty-leg flights), a fitness center / gym (memberships, personal training, group classes, merchandise, and a supplement bar), plus a 5-year cigar lounge and a household budget — so you can see how the tool reads whatever you'll actually upload.

›Want the exact format to start from? Load any sample, then Export → Full dataset to download it as a template (CSV, TSV, JSON, or Excel).

›SIMPLE VIEW — feeling like it's a lot of numbers? Tap the Simple view toggle at the right end of the tab bar: five big numbers and the story in three plain sentences, nothing else. Tap any number to see the rows behind it, and "See the full picture" brings the full dashboard back. Your choice is remembered.

›An Executive summary sits at the top of the dashboard as soon as data loads — no export needed to see it.

›KPI cards cover total revenue, records, average, active count of your main grouping (clients, patients, categories — whatever the data is), latest month, and best month, plus a second row for Month-over-month, Quarter-over-quarter, and Year-over-year (YoY needs 24+ months of history to appear).

›PlainSight names your grouping after what it actually is, rather than calling everything a client — patients for a dental export, properties for rentals, trucks for freight, wells for oil & gas. So a production file shows Top wells and Active wells, not Top clients.

›If your data also has an operator or company column separate from your main grouping, a Wells / Operators toggle appears on the Top panel. Operators rolls revenue up by operating company — Chevron, Hess, EOG — which is the view a services firm or a consultant building a competitor analysis wants, since the operator is the client and the wells are the work. The toggle only appears when that column exists.

›What your numbers are telling you sits on the Dashboard, above the executive summary — actions read straight from your file with no AI call, no industry pick, and no button. It flags client concentration (anything over 30% from one client is called out as risk, with the dollar exposure), month-over-month momentum, a long tail of low-value clients, how far below your best month you're running, and what a 10% lift in average record size would be worth. Every item carries a projected impact and a timeframe. The Benchmarks & actions tab goes further by comparing you to your industry; this is the floor, not the ceiling.

›Slicing by client, well, or whatever your grouping is: a handful show as chips, and past eight you get a searchable multi-select picker — type to filter, tick several at once. Nothing is ever hidden from you; earlier versions capped the chips at eight and dropped the slicer entirely past forty.

›OIL & GAS REVENUE — A production file contains volumes, not revenue, so PlainSight doesn't pretend otherwise. Load one in Revenue mode and a Well economics panel appears: revenue = volume × realized price × your net revenue interest. Set oil and gas prices (or tap Use live prices to pull the current WTI and Henry Hub from the Energy prices feed), add a basin differential if your crude sells at a discount to the benchmark, and set your NRI — the share you keep after royalties. Everything on the tab — totals, charts, top wells, insights — is then computed from real dollars.

›Lease operating expense is optional. Leave it blank and you get gross revenue, royalty, net revenue, and revenue per BOE. Enter an LOE in $/BOE and three more figures appear: operating cost, netback, and operating margin — along with a working-interest field, because costs are borne at your WI while revenue is received at your NRI, and those are usually different numbers. BOE uses the standard 6 Mcf = 1 bbl equivalence.

›If your file already has a money column — an oilfield services or consulting invoice export, say — none of this appears; it's already revenue and the normal dashboard applies. Try the 🛠️ Oilfield services sample for that case and the ⛽ Oil & gas well sample for the operator case.

›The revenue chart has two groups. TREND views: Bar, Line, Area, Cumulative (a running total), and Combo (monthly columns on the left axis plus the cumulative total on a right axis — a true dual-axis view). BREAKDOWN views: Pie, Donut, Ranked (top entities as horizontal bars), and Pareto (the 80/20 — entity bars with a cumulative-share line and a dashed 80% guide, so you can see how few entities make up most of the total). A ★ Suggested chip points to the best chart for your data (e.g. Pareto/Pie when revenue is concentrated, Combo for a long growing series). Line and Area also offer From-zero or Zoom-to-data scaling, and long multi-year series render as dense, contiguous columns so they stay readable on a phone.

›Slicers let you filter by period, by your main grouping column (whatever it is — client, patient, department, category…), or any detected category — including a location column if one's present, which also unlocks a Revenue by location panel.

›Deeper numbers shows median, typical range, and average monthly growth. The box-and-whiskers plot shows how transaction sizes spread — a tight box means consistent pricing, a wide one means variable job mix, with its own insights folded into What the numbers say. Toggle it by month, by your grouping (client, artist, provider…), or — when a service/category column is detected — by service, and tap any column to see its rows.

›A Recommended actions preview appears once you've generated benchmarks, with a link to the full list.

›In Benchmarks & actions — or straight from the dashboard toolbar, where your industry shows next to the view toggle — set your industry (type anything — it fuzzy-matches), your company size (Solo / Small / Medium / Large — this changes what "typical" means), and optionally your area for local comparisons, then generate a scorecard and prioritized action steps.

›Each recommended action includes a projected impact — a specific number computed from your own data, like the added clients needed to fix concentration or the monthly revenue a pricing change would add — plus a timeframe for when to expect to see it move. In Operations mode, impact is always a rate or percentage, never a dollar figure, since commodity price and cost aren't known here.

›An Executive summary specific to that run appears once results are in, reflecting your actual selections and findings — it reads differently depending on what you chose.

›A Performance at a glance chart sits above the detailed scorecard: one bar per metric showing how far you sit from the middle of each typical range. Bars to the right are healthy, bars to the left need attention, and a metric sitting inside its typical band shows a marker on the line. Metrics where lower is better (concentration, volatility) are flipped so the direction always means the same thing.

›If the AI is unavailable, a built-in rules engine computes benchmarks — and projected impact — locally from your data, so you always get results.

›AI NARRATIVE (Pro) — below the benchmark controls, Generate narrative writes a short written brief on your data: two or three tight paragraphs that lead with the most important takeaway, cite your actual figures, weave in the benchmarks when you've run them, and close with the one move that matters most. It's the "hand it to a partner or lender" version of your dashboard, and it also appears as an Analyst narrative section in the downloadable PDF and Word reports. In Trading mode the same button writes a coaching review of your journal instead — edge, expectancy, drawdown, and where the leak is. A 🔊 Listen button in the write-up reads it aloud using your browser's built-in voice (free, on-device, no audio is sent anywhere — and it automatically picks the most natural voice installed on your device); press ⏹ Stop to end it. Free users see a clearly labeled example so you can judge it before upgrading — with the same Listen button.

›Labels adapt to your data everywhere — a dental export says patients, a grocery export says departments, a tattoo shop says artists, a barber shop says barbers, a salon says stylists, a nail salon says technicians, a med spa says providers, a household budget says categories, a stock log says symbols, a yacht-charter export says vessels, a bar or lounge says revenue streams. PlainSight reads the grouping column and industry and never forces the word "client" onto data where it doesn't belong. Household budgets go further: spending instead of revenue, and household guidelines instead of business norms.

›Benchmarks & actions and Sector news apply to Revenue and Operations data; Trading mode has its own built-in actions on the dashboard instead. In Operations, benchmarking follows your well selection: one well is compared as itself, while a group is compared as its average well (per-well IP, peak, and decline averaged, with ranges) — see the Oil & gas section for the details.

›Sector news does more than list headlines. Each recent article is tagged by theme (Regulation, Pricing, Demand, Technology, Costs, Competition, Labor) and whether it's a tailwind or headwind for you. A News signal map at the top shows the balance at a glance — whether your sector's recent news is net-positive or net-negative and which themes are driving it — and a What this means for you panel turns the headlines into 1–3 recommended actions, each tied to a specific article with a projected impact and timeframe. If you've already run benchmarks, the search is weighted toward what those flagged.

›Not connected to AI yet? Tap Preview with sample news to see the full layout — signal map, tagged headlines, and news-driven actions — populated with realistic, industry-matched sample coverage. It's clearly labeled as a sample, never live, and uses the exact same rendering as the real feature, so what you see in the preview is what you get once AI is connected.

›Ask your data sits above the tabs — ask plain questions about your numbers in Revenue or Operations mode. Pro includes 300 questions a month, counted on the server; a $9 top-up adds 500 more that don't expire, and the monthly allowance is always spent first so the perishable one goes before the one you paid for. It is not available on Free — reading your own figures is the paid half of the assistant. If a question fails, it isn't counted. If AI is unavailable, common questions (best/worst month, growth, concentration) still get an instant local answer. Only anonymized summary totals are sent — never raw rows, and the server refuses a payload that looks like rows even if the app sent one by mistake.

›Options and futures are read from what the file SAYS, not guessed from the ticker: if your export has an instrument column (thinkorswim's Type, for instance), that decides it. Options pair against their own contract — expiry, strike and right — never against the underlying shares, and carry the 100x multiplier. A plain ticker is always treated as stock, because futures roots collide with real companies: CL is Colgate as well as crude oil, ZM is Zoom as well as soybean meal. Futures need to be written the way your broker writes them (/CL, CLM24, CL 03-26). Two sample shapes are provided: "Trading journal" is one row per completed trade with a P/L column, and "Broker export (raw fills)" is separate buy and sell fills with no P/L at all — the shape most brokers actually give you, which PlainSight pairs into round trips itself. If you're not sure your export will work, open the broker sample first; it's the same shape. Load a trade log or broker export and PlainSight switches to Trading mode automatically — or press Trading in the top-bar toggle, which offers all three views once a file is loaded. If your journal uses column names PlainSight doesn't recognize it lands in Revenue — pressing Trading is how you correct that, and it tells you exactly which columns it needs if it still can't read them. Everything is analyzed in your browser and is never stored or mixed into sample data.

›Two file shapes work. A journal export already has one row per trade with a P/L column — that's used directly. A raw broker export lists separate buy and sell fills with no P/L; PlainSight pairs them FIFO into round trips per symbol, nets out commissions and fees, handles shorts and same-day flips, and excludes any still-open position from the stats (it tells you which ones). Options get the 100× contract multiplier; futures are scaled by each contract's point value (ES $50, MES $5, CL $1,000, GC $100, and ~40 more, with expiry codes like ESZ5 or /MESH6 resolved to their root). THE ROAD RECAP — set your industry to comedy/performer and the Revenue dashboard reads a gig log like a session recap: last month, best room and its share, rooms worked, cadence, average take, and the 90-day trend, all computed on your device (numbers are Pro; a free 'Preview with sample numbers' shows the shape first, and the recap links across to The Circuit, the comedy side's weekly industry brief). A performer's whole stack fits one book: load a gig log, then ＋ Add data with the podcast or merch export as second and third sources — the Road Recap reads the split (≈ by source) and the source panel slices it. Journal-shape logs are read for more than P/L too: a Fees/Commission column unlocks the three-way cost views, and an Instrument column (Stock / Option / Crypto / Futures) feeds the Instrument chips — both picked up automatically when your log carries them.

›The dashboard leads with six KPIs measured together — Net P&L, Win rate (and % of days green), Profit factor, Expectancy per trade, Max drawdown, and Avg R-multiple — because no single one tells the whole story. A $/R toggle recalculates everything in R-multiples; if your log has no risk column, 1R is estimated from your average loss and labeled as such. A cost toggle shows the book After costs, Before costs, or Costs only (the fee ledger alone) — it opens on Before costs, and the Costs line always keeps both numbers in frame. A Grade card marks the filtered slice A+ through F on a four-part rubric shown in its ⓘ — no grade under 30 trades. Custom date ranges (set the date boxes, or simply drag across a stretch of the equity curve — release, and the filter snaps to those trades' dates), year-jump calendar arrows, and day rows with entry → exit prices round out the journal.

›Below that: an equity curve with drawdown shaded from each peak, an interactive P&L calendar (tap any day to see its trades), and breakdowns by setup, time of day, day of week, long vs short, symbol, and holding time. Any dimension your file doesn't include shows a short hint on how to add it rather than disappearing.

›Recommended actions carry the same projected impact + timeframe as the rest of PlainSight — e.g. "Pause the VWAP Reversal setup — +$5,795 kept over the last 8 months" — and only fire on slices with enough trades to matter. Under ~30 trades, the one recommendation is to keep logging until the sample is readable. When statement fees are on the book, one more card can appear — Where the toll concentrates: your average fee measured against your average pre-fee win, by hold time and by instrument (options vs shares). It only speaks on real contrast — the worst pocket paying at least 10% of a typical win and at least twice the calmest one's share, with enough trades in each to mean it. If you don't see it, your costs are spread evenly, and the Costs line under the KPIs already tells that story.

›Filters at the top scope the view by date range, setup, account, year, and instrument (Options / Shares / Futures / Crypto chips appear whenever the book holds more than one; crypto is recognized when the export's instrument column declares it); actions always score on your full history so they stay stable while you slice. Trading mode is self-contained — it doesn't use the industry Benchmarks, Sector news, or Ask-your-data tabs, which are built for revenue and production data.

›MARKETS & PREP — Trading mode has a second tab: a desk for the session either side of the bell. Last session recap reads your own log — net P&L, hit rate, best and worst trade — and flags the habits that cost money: a position that jumped past your usual size, a size increase straight after a loss (the revenge pattern), or a trade count well above your norm. Prep for the next session turns your own stats into a short pre-open checklist. THIS WEEK VS LAST sits between them: net P&L, trades, hit rate, profit factor and average loss for this week against the week before. It matches LIKE FOR LIKE — the first N trading days of each week, where N is however many the current week has so far — so a Tuesday is never compared against a full five sessions, and the card says so when last week has been trimmed. The green tick notes only ever cover things the trader decides: fewer trades when they were overtrading, no size increase after a loss, a smaller average loss or worst loss, and a better profit factor only when it was earned by cutting losers smaller. P&L is reported and never congratulated — a green week can come from luck or from bigger size, so if someone asks why a profitable week got no praise, that is the answer and it is deliberate. Needs at least three trades in each week or the card does not appear at all. Red-folder calendar lists the high-impact US releases ahead; NFP (first Friday) and weekly jobless claims are computed exactly, FOMC comes from the Fed's published schedule, and CPI/PCE show their usual window — always confirm the exact day on the official calendar. All of that is computed on your device and free on every plan. The Market board pulls a snapshot via web search on Pro (Free shows the same layout with sample values): indices, index futures, energy & metals (WTI, Brent, natural gas, gold, silver), crypto, the seven major USD FX pairs, movers and gappers, and sector leaders vs laggards. Tap ⚙ Customize to edit any row — swap the FX pairs, add contracts you actually trade, or clear a row to hide that card. It's also your speed dial: fewer symbols means a faster snapshot, and results are cached for five minutes so Refresh is instant. My watchlist is separate and yours to edit. Building one needs a free account — the list is tied to your account so it follows you between devices instead of living in one browser, and it's what personalizes the Market Watch briefing. Check my list looks those symbols up on their own request so a long list can never slow the main board, and live quotes for it are Pro. Quick links jump you out to the rest of the desk — CNBC for headlines, Nasdaq trade halts for live halt codes and resumption times, Finviz for the sector map, TradingView for charts, Market Chameleon for options and earnings analytics, Forex Factory for the pre-open calendar with impact ratings, consensus and prior alongside each release (including the non-US events the US calendars omit), and the CME and BLS calendars for exact release dates (the red-folder note links straight to BLS and the Fed so you can confirm a date in one tap). WHAT HAPPENS TO MATERIAL PASTED INTO THE PUNCH-UP DESK: it leaves the browser once, to Anthropic's API, and comes back edited. PlainSight keeps a count of drafts run and never the text itself — the only thing written is a per-tier counter. Anthropic's commercial terms bar training on customer content, and they delete API inputs and outputs within 30 days; that is their published policy, linked on the page, not a PlainSight claim. The desk has two modes: punch-up EDITS material the comic already has, and the other mode WRITES from a few details — so a flat 'we never write material' would be false and support must not say it. YOUR ACT (punch-up desk): four optional lines a comic writes once — their angle, their register, the subjects they own, and what they never do. The desk edits against those instead of guessing, and a margin note that cites one of them ('cut: this is the kind of family material you said you don't do') is craft feedback rather than 'tightened'. It is stored ON THEIR DEVICE only, never on the account and never on our servers — the list of subjects somebody refuses to touch is more personal than the bit. It travels with the draft request that needs it and nowhere else, and one tap clears it. Deliberately NOT called a persona: that word already means the revenue/operations/trading audience mode elsewhere in the product. DECLINE CURVE METHOD, for operators who ask: modified hyperbolic Arps with a terminal decline switch at 6%/yr effective, and P10/P50/P90 read off the aligned well matrix rather than off a fitted average. Market figures are indicative and may be delayed — verify with your broker before acting.

›MARKET WATCH NEWSLETTER (Pro) — the desk, delivered. One email each weekday at 8:00 AM ET, before the open: where the prior session closed and what drove it, how Asia, Europe and the futures traded overnight, the red-folder calendar of scheduled high-impact releases, what's on today's docket, and the individual names with a genuine scheduled catalyst (an earnings date, a guidance change, an unusual gap) with the reason stated as fact. If you've saved symbols to My watchlist, the brief adds a short section reporting what happened to those specific names — reporting only, never a view on what to do about them. Subscribe from the bottom of Markets & prep; the address is yours to choose and is used for nothing else. Every issue carries a one-click unsubscribe that works without signing in. IMPORTANT: this is market commentary, not investment advice. PlainSight is not a broker or investment adviser, makes no buy or sell recommendations, and states no price targets. Figures come from public sources, may be delayed or wrong, and must be confirmed with your broker before you act.

›THREE MODES — PlainSight picks the right view for your data automatically: Revenue mode for sales, invoices, and payments (any business); Operations mode for oil & gas production volumes; and Trading mode for trade logs and broker exports. A top-bar toggle shows all three whenever a file is loaded, whichever one PlainSight picked — so if it read your file the wrong way, you correct it in one tap rather than re-exporting. Next to it sits your industry, which is a separate choice: the mode picks the dashboard, the industry picks the benchmarks and sector news.

›OIL & GAS vs. OILFIELD SERVICES — these are different businesses and PlainSight treats them differently. An OPERATOR's production file contains volumes (oil, gas, water — no dollars), so it lands in Operations mode with decline curves, type curves, and well economics computed from your price assumptions. An OILFIELD SERVICES or consulting company invoices operators in dollars — that's ordinary revenue data, so it lands in Revenue mode with operators as your clients. Try the ⛽ Oil & gas well sample for the operator case and the 🛠️ Oilfield services sample for the service-company case.

›Upload a production file with oil/gas/water or pressure columns and PlainSight switches to Operations mode automatically — or press Operations in the top-bar toggle, which offers all three views once a file is loaded.

›See rates (bbl/d, Mcf/d), pressures (psi), decline curves, and a Combined dual-axis view (rates left, pressure right). Switch volume units between bbl, Mbbl, and MMbbl.

›If your file has a well-type column (Horizontal/Vertical/Injection), a slicer appears above the dashboard.

›SELECTING WELLS — With more than one well, a Well selector focuses the dashboard. For a few wells it shows as chips; for many, a compact dropdown you can type into (search by name, number, or type). Select several at once with the checkboxes, or use the ‹ › stepper to page through them one at a time. Clicking wells on the Basin map toggles them in or out of the same set.

›ONE WELL vs. A GROUP — With a single well, KPI cards show IP (early rate), peak, latest, and cumulative per stream, plus TD/MD/KB depths and an onshore/offshore badge. Select several wells and the numbers change meaning, so PlainSight labels them explicitly: cumulative volumes are summed (combined cum, with an avg/well figure beside it) because volume is additive, while IP and depth become per-well averages — each well's own value, then averaged — shown with the min–max range across the group. That range is the point: it tells you whether one strong well is lifting the average, and what a typical lateral depth and well quality look like in that area. Rate charts add the wells' daily rates and average their pressures (summing pressures isn't physical).

›The Executive summary at the top reflects whatever you've selected — one well's IP, peak, decline, and completion, or, for a group, the average IP per well with its range, combined production, typical depth, and average completion, calling out when the spread is wide enough that one well is carrying the group.

›DECLINE CURVE ANALYSIS — With one well selected, the Decline snapshot fits the MODIFIED Arps model to your rate history and reports what a reservoir engineer works in: b-factor (with the model named — exponential, harmonic, or hyperbolic), initial decline Di as an effective annual rate, current decline (which flattens as the well ages), initial rate qi at peak, EUR split into produced plus remaining, and remaining life to the economic limit. Every fit carries an R², and below 0.75 PlainSight says the fit is loose and to treat EUR as indicative.

›Two things about the math that matter. Fitting starts at the PEAK, not the first reading, so flowback and ramp-up don't drag the curve. And the model is modified Arps, not pure Arps: hyperbolic decline runs until the instantaneous decline falls to a terminal minimum (6%/yr effective by default), then switches to exponential. That switch isn't decoration — pure hyperbolic with b greater than 1 never converges, so a b=1.5 well would keep 'producing' forever and its EUR would be meaningless. Real b-factors for unconventional wells sit in that range, which is exactly why the industry uses the terminal switch.

›TYPE CURVES — Select several wells and you get a type curve, built the way the industry builds them: each well is time-zeroed at its own peak (wells come online in different years, so calendar time tells you nothing about shape), the aligned rates are averaged step by step, and THAT averaged profile is fitted. The result is the average well, not the summed field — summing would give a curve whose b-factor describes no well that exists. You also get P10/P50/P90 at peak, the b-factor and qi ranges across the group, and a warning when the best well is 2x the weakest, since that means the average is being pulled by the top end and may not represent the group. When every selected well has a lateral length, you can normalize to rate per 1,000 ft so a 10,000 ft lateral compares fairly against a 5,000 ft one. Each well is also fitted individually, which is what the per-well parameter ranges are built from.

›EXPORTING THE MATH — Show the math opens the constants panel: the equation with your actual numbers substituted in, and every constant editable. Change qi, b, Di, the terminal decline Dmin, or the economic limit and the model re-runs immediately — equation, EUR, remaining reserves, remaining life, and the terminal-switch time all follow, and so does anything you download afterwards. Edited fields are marked with a dot; Reset to fitted puts them back. Constants belong to the well they were tuned on, so switching wells clears them rather than carrying a hand-typed b onto someone else's curve.

›One thing PlainSight is strict about: the moment you edit a constant, R² is RECOMPUTED against your real rate history. The fitted R² described the fitted curve — reporting it next to hand-typed numbers would be a lie. So if you widen b past what the data supports, you'll watch R² fall, which is the point. Exports carry your constants, tag each one as fitted or USER-SET, flag the file as manually adjusted, and list the original fitted values underneath, so whoever opens it next can tell what came from the data and what came from you.

›Export forecast (.csv) writes a month-by-month rate, volume, and cumulative table out to the economic limit, headed by the full equation set, every constant with its provenance, the terminal-switch time, and R² — so it loads into Aries, PHDWin, ComboCurve, Mosaic, Excel, or a Python notebook, and another engineer can audit or re-run it. If a curve is still above its economic limit at 50 years, the forecast is cut there and both the panel and the export say so — EUR is then a floor, not the whole curve. On a group you can also export the type curve and a per-well parameter table (qi, b, Di nominal and effective, current decline, model, R², cum, EUR, remaining, remaining life, switch time) for portfolio work.

›BENCHMARKS — Compare against a typical well, this basin, or (with 2+ wells uploaded) your own other wells' real averages. With several wells selected, PlainSight benchmarks the average well in the group — per-well IP, peak, and decline averaged, with ranges — rather than a summed curve, which would describe a well that doesn't exist. GOR and water cut stay as combined ratios, which are validly volume-weighted. Peer comparison excludes every well you've selected, so you're never compared against yourself.

›Free & Pro: "Benchmark this well" runs free on your device with the built-in rules engine — a real read of your own wells, nothing sent. Pro turns that same button into a live AI reservoir analysis of your loaded wells. On Free (and Report pack) a "Preview AI-enhanced sample" button shows an example of what that AI read looks like; on Pro the preview button disappears because the real thing is one click away.

›WELLBORE & COMPLETION — These are per-well by nature; with a group selected they show the first selected well and say so. Completion details (formation, lateral length, frac stages, stage spacing, proppant, proppant type, proppant per stage, proppant intensity in lb/ft, fluid volume, fluid per stage, spud/completion dates) come from an upload — a standalone file or a workbook tab, matched to wells by name.

›The wellbore panel draws a geological cross-section: the vertical bore kicking off into a horizontal lateral, with frac stages marked along it from your real stage count and light proppant scatter around each. Injection wells draw dashed; vertical wells stop at TD. Without formation tops, the layers are that basin's actual stratigraphy in correct order (a Permian well shows Bone Spring, Wolfcamp A/B/C; a Bakken well shows Upper/Middle/Lower Bakken, Three Forks) at proportional depths — real names and order, estimated depths. Your landing zone is outlined in gold and the lateral lands in it.

›FORMATION TOPS — Upload your surveyed picks and the wellbore switches to your real geology: layers drawn at their actual depths, badged SURVEYED TOPS, and the disclaimer changes accordingly. Use a long-format file, one row per formation per well: Well, Formation, Top Depth, and optionally Base Depth (inferred from the next top if omitted). Column names are flexible — Zone/Marker/Horizon/Pick and MD/TVD all work. The Download template button beside the wellbore gives you the exact shape pre-filled with your own well names. The oil & gas sample loads representative tops so you can see it working before you have your own.

›EXPORTS — The Export menu names exactly what you'll get before you click: which wells it covers and which tab it's from. Reports are tab-aware — export from Dashboard and you get production, rates, per-well breakdown, economics, completion, tops and basin context; from Benchmarks & actions you get benchmarks and recommended actions; from Sector news you get prices and headlines. There's also an Everything option that stitches all three into one document. Multi-well selections export as the whole group, including a well-by-well table. Reports download as HTML — open one and use your browser's Print → Save as PDF; that's deliberate, since there's no PDF engine running in your browser.

›BASIN MAP — Plots your wells and overlays sample national wellheads to gauge area activity. Three views: Full US (clean and uncrowded), Basin (zoom to fit), and Selected well (zoom tight on the well you're focused on). Hover a well for its name, type, and IP rate, or a basin for its rig count and how many of your wells sit inside it. Click your wells to filter the dashboard; double-click one to jump straight to its wellbore and completion detail. Click any basin to compare against it.

›RIG COUNTS — Default to the real Baker Hughes weekly figures, baked in. Alongside the current count you get week-over-week, month-over-month, and year-over-year change plus a 26-week sparkline, for the U.S. total and for each basin. Download the latest Baker Hughes file from the link provided and upload it to refresh: PlainSight reads the weekly detail sheet and rebuilds the counts, the deltas, and the trend lines together. Because rig counts are public information, this refresh stays open on every plan — free included. Live EIA auto-updates and licensed permitted-well data are coming soon.

›SECTOR NEWS — Opens with an Energy prices strip (WTI, Brent, Henry Hub, front-month futures with day-over-day change) pulled from the same live web search as the news; sample values show in preview mode. Below it, recent headlines are tagged by theme and tailwind/headwind, with a signal map and recommended actions.

›Free & Pro: on Free (and Report pack), "Get sector news" shows sample headlines (the baseline), and a "Preview AI-enhanced sample" button adds the recommended-actions layer so you can see what Pro produces. Pro's "Get sector news" runs a real live web search for your sector and writes the full digest; the preview button goes away.

›WHERE TO SUBSCRIBE — Both briefings are listed on the app's home screen under "✉️ Briefings by email", just below pricing — those cards load the right sample and take you straight to the signup. The Oil & Gas Weekly signup itself sits at the bottom of this Sector news tab. It also appears in REVENUE mode's Sector news tab whenever your industry reads as energy — oilfield services, consulting, wireline, frac and the like — because those files are invoices rather than production volumes and land in the revenue dashboard, but the people reading them want the same brief. There are direct buttons on the home page too.

›OIL & GAS WEEKLY NEWSLETTER (Pro) — Written for people who work in the patch rather than for traders. One email every Monday at 6:00 AM ET covering the week's most consequential sector news with the specific figures, an explicit tailwinds and headwinds read, the latest Baker Hughes rig count with week- and year-over-year change and the top basins, current energy benchmarks, the week's scheduled energy events (EIA petroleum status Wednesday, Baker Hughes Friday), and recommended actions grouped by discipline — Operations & Production, Drilling & Completions, Land & Regulatory, Commercial & Marketing, Finance & Planning, and HSE. Actions are operational and commercial, tied to something that actually happened that week; they are never securities recommendations. Monday morning is deliberate: Baker Hughes publishes Friday afternoon, so the weekend leaves room for the rig refresh to land before the send. If the stored rig snapshot is more than 10 days old the rig section is omitted rather than quoting stale counts, and the operator is emailed about it. Subscribe from the bottom of the Sector news tab. Industry commentary, not investment, engineering, or legal advice — validate against your own data and your regulator's current guidance.

›For paper-ledger businesses: upload up to 5 photos (a ledger, receipts, even a handwritten page) and AI vision reads every legible transaction it finds, combining multiple photos into one dataset.

›Nothing loads automatically — you review an editable table first. Fix any date, description, or amount before committing; fields the AI wasn't confident about are flagged for a second look.

›This is the one feature where content leaves the device, since reading a photo requires AI vision. It counts toward the same monthly AI cap as Ask your data and Benchmarks.

›Downloads and exports are a Pro (or Report pack) feature — the free plan is view-only on sample data. Once you're on a paid plan, Revenue mode's Export menu is grouped so it's clear at a glance. FULL REPORT is the complete branded deliverable — cover, executive summary, key metrics, analyst narrative, every chart, industry benchmarks and recommended actions — in three formats: PDF (print / save as PDF), PowerPoint (.pptx, with a short observation on each chart slide), or Word (.doc). CHARTS (IMAGES) exports chart PNGs — either just the view you're looking at (This view), or every view (All views: Bar, Cumulative, Combo, Pie, Ranked, Pareto and the box plot) captured automatically and stacked into one image, even the views you don't currently have open. DATA covers the .txt report, .csv summary, the Full dataset in CSV/TSV/JSON/Excel, well economics (.csv) for oil & gas, and a copy-ready email summary. Each chart also has a ⤓ PNG button for a one-tap image of just that chart.

›Operations mode has its own Print/PDF (a production-rate chart plus stream and completion tables), a Download dataset button, and the DCA exports (forecast, type curve, per-well parameters). Trading mode has a Download dataset button too — including the round-trip trades PlainSight paired from a raw broker export.

›The My Recommended Actions tab gathers every action PlainSight flags — from your data, benchmarks, and sector news — into one checkable list you can export to CSV. And a Read aloud button reads a spoken summary of whatever tab you're on (dashboard, benchmarks, or sector news) aloud in English using your device's built-in voice — free, on-device, no AI cost. (The 100+ languages feature translates the on-screen text; the spoken summary is in English. If you'd like it read aloud in another language, your browser's own built-in Read Aloud will speak the translated page.)

›DATA SHEET (also called the data table or data drawer) — two ways in, and both work. There IS a visible 📄 Data sheet button in the dashboard toolbar for the whole dataset, and it's in ⌕ Jump to… as well. You can also go straight from any figure: on a mouse or trackpad, a SINGLE CLICK on a bar, line point, pie slice, KPI card or location row opens the drawer immediately. On a phone or tablet the first tap shows that element's tooltip and a SECOND tap opens the drawer — deliberately, because a single tap would make it impossible to read a tooltip without the sheet flying open. Either way you get the exact rows behind that number: sortable, read-only, always matching your uploaded source. Nothing in PlainSight is a black box — every figure traces back to the rows that produced it, which is what you need if someone asks you to prove a number. The guided tour demonstrates this in all three modes. Viewing is free for everyone; downloading the sheet as CSV is Pro (or Report pack).

›ON A PHONE — the dashboards work in portrait, but charts get roughly twice the width in landscape, so PlainSight shows a one-time "turn your phone sideways" hint the first time you open a chart on a narrow screen. It's a suggestion, not a requirement: rotation is never forced, because locking orientation doesn't work at all in iOS Safari and overriding someone's rotation lock is presumptuous. Dismiss it once and it won't come back.

›The Print / Save as PDF report is a polished, branded deliverable: a cover with your logo and contact details, a clickable table of contents, an executive summary, a key-metrics table with a transaction-spread box plot, the revenue trend, top entities with a concentration Pareto (the 80/20 view), plain-language findings, a sector-signals section, industry benchmarks (when run), and a methodology & disclaimer page — the kind of document you can hand a client or lender. The exact same report is also available as a PowerPoint deck (.pptx — one section per slide, with a short observation on each chart) and as an editable Word document (.doc), all under the Full report group of the Export menu. The on-screen executive summary is also clickable: each line drills to the exact rows behind it.

›Every export leads with an auto-generated executive summary, and the .txt report includes each action's projected impact and timeframe when you've generated benchmarks.

›SAVED REPORTS — once data is loaded, 💾 Save report stores that analysis on your own device so you can close the tab and pick it up later with every filter still available. Saved reports appear under 💾 Saved reports in the upload area, listed with row count, industry, and date. They are held in your browser's local database (IndexedDB) — nothing is uploaded, we hold no copy, and they stay on that device until you delete them or clear your browser data. Saving is a Pro (or Report pack) feature; note that saved reports don't sync between devices or browsers, and private/incognito windows usually can't save at all. To move a report to another device, use 📲 Send to my phone — see DEVICE HANDOFF.

›If a direct download or print dialog is blocked — common inside an in-app browser — PDF exports instead download as an HTML file you can open and print from any browser, and other exports fall back to a copyable text panel.

›DEVICE HANDOFF — 📲 Send to my phone moves the report you're looking at onto another device without re-uploading any files. The browser compresses and encrypts the report on your machine, uploads only the encrypted result, and shows a QR code (and a copyable link). Scan it with the other device's camera, sign in there with the SAME account, and the book opens with its sources and industry intact (filter selections start fresh), then saves itself to that device's Saved reports so it survives closing the browser. The decryption key travels in the link's # fragment, which browsers never send to a server — so PlainSight relays the transfer without being able to read it. A handoff opens once, expires after 15 minutes, and is refused for anyone signed in as a different account. Available to any signed-in user, not just Pro. Requires being signed in on both devices — if the receiving device isn't signed in yet, you'll be prompted, and the report finishes loading on its own after you sign in (even when sign-in bounces through Google). Re-scanning the same code also works: a signed-out attempt never uses up the handoff.

›THE ASSISTANT, ACROSS THE PLANS — the free navigator ("Ask for what you came for") is open to everyone including signed-out visitors, ten questions a day, and never sees your data. The analyst ("Ask your data") reads the summary totals of a loaded file and is Pro: 300 questions a month. A one-time $9 top-up adds 500 that do not expire and can be bought on any plan. A Report pack does NOT include analyst questions — it is a one-time purchase, not a subscription, so pack holders get the free navigator like everyone else. Counting happens on the server, so it survives a refresh.

›Free explores every industry on sample data — no account needed — and includes ONE upload of your own file: create a free account and PlainSight builds your full on-screen dashboard with the on-device rules engine (KPIs, charts, straight-from-data recommended actions), saved to your account so you can come back to it. You can also download the one sample for your chosen industry as a format template. Reserved for Pro: viewing or downloading the summary/export report, any export of your own analysis (datasets, Data Sheets, chart images), and all AI features. Refreshing the Baker Hughes rig counts is public information and stays open on every plan.

›Pro ($19.99/mo or $99/year) unlocks unlimited uploads, every download and export, the full summary report (view + download), the combined book (add multiple accounts/locations/files into one view with a per-source breakdown), plus the live-AI layer: live sector news, AI benchmark commentary, the AI narrative write-up, Ask-your-data, and ledger-photo reading. Saving reports to your account is included free once you've signed in.

›Report pack ($12.99, one-time — no account needed) lets you upload your data once and download a full report on that device, computed by the built-in rules engine. It's the no-subscription option; it doesn't include the live-AI features.

›PlainSight works fully without AI — every dashboard, chart, benchmark, and set of recommended actions is computed on your device by a built-in rules engine. The AI layer (Pro) adds live sector news, richer benchmark commentary, free-form Ask-your-data answers, and ledger-photo reading on top of that. On the free plan the news tab still previews sample coverage and offers one-tap live search links, and benchmarks are computed locally — so every tab stays useful at no cost.

›Your uploaded files, pasted data, and spreadsheets are processed entirely on your device and are never uploaded or stored.

›Optional AI features (benchmarks, news, Ask your data) send only anonymized summary totals — monthly figures and "Client A: 34%" style shares — never raw rows, names, or file contents.

›The one exception is ledger photos: reading a photo requires sending the image itself to AI vision for one-time processing. Nothing else about your business is shared as part of that, and it's entirely opt-in.

›Saved reports stay on your device too. Choosing 💾 Save report writes that analysis to your browser's own local database (IndexedDB) — it is never transmitted to us, we hold no copy, and it remains only until you delete it in the app or clear your browser data. Favorites and interface preferences are stored the same way. The one exception is 📲 Send to my phone: to move a report between your own devices we relay it through our server, but it is encrypted in your browser first and the key is never sent to us — it travels in the link's # fragment, which browsers don't transmit. We hold ciphertext we cannot read, for at most 15 minutes, and it is deleted the moment your other device opens it. Full details are in our Privacy Policy.