Docs › Desktop app
The desktop app
The Insight FastPass desktop app is the report builder on your own computer. It reads databases on your network directly, with no gateway to install, keeps a copy of what it reads, and saves each report with its data in one .ifpr file you can open again later, on the desktop app or the website. Signing in is optional. Signed in to a Plus or Ultimate workspace, you also get that workspace's library on your computer and can publish to the workspace.
Plan availability Anyone can use the desktop app for reports saved as files, without an account. A library, and publishing to a workspace, need you signed in as a member of a Plus or Ultimate workspace.
What it does that the website doesn't#
| Website | Desktop app | |
|---|---|---|
| Reach a database inside your network | Through the Data sources add-on and a gateway (see Install and pair a gateway) | Directly, from your computer, with your own sign-in. No add-on, no gateway |
| Where your data is kept | In your workspace, by plan | In the .ifpr file you save, or in your library on your computer or a network folder |
| How many datasets and reports | Your plan's limits | No report count. A library has no dataset or report limits |
| Work offline | No | Yes, once installed. Signing in and publishing need the internet |
The website is still where you manage your workspace: people, roles, billing, security, data sources and Ask FastPass's settings. The desktop app opens the website for those. Ask FastPass itself works on reports in the app too (see Ask FastPass about a report on your computer).
Install it#
| System | Installer |
|---|---|
| Windows 10 and 11 | An .exe setup, which installs for your user without administrator rights, or an .msi for installing on every account of a computer (needs an administrator, and suits IT deployment). Both use the Microsoft Edge WebView2 runtime, which Windows 11 already has; the setup fetches it on Windows 10 if needed |
| macOS 11 or later | A .dmg: open it and drag Insight FastPass to Applications |
| Linux (Ubuntu 22.04 or later, and similar) | A .deb, an .rpm or an AppImage |
Installing it registers .ifpr files with the app, so double-clicking one opens it.
Some databases need their own software on your computer as well:
- Oracle needs Oracle Instant Client 19 or later. Install it and put its folder on your
PATH(Windows),DYLD_LIBRARY_PATH(macOS) orLD_LIBRARY_PATH(Linux). Without it the app still works; testing an Oracle connection says the client is missing. - ODBC uses the ODBC drivers installed on your computer. Set up a data source name (DSN) in your system's ODBC administrator, or use a connection string. On macOS and Linux, install unixODBC and your database's driver.
- SQL Server, PostgreSQL and MySQL need nothing extra.
The start page#
The app has one layout throughout: a sidebar on the left, a bar across the top, and the page you're on beside them. The sidebar's Desktop section has Home (this start page), New report, Open… and Settings. At the bottom of the sidebar, Support ↗, Docs ↗ and Community ↗ open the website in your browser. Once you're signed in, your workspace's sections follow it (see Your library). At the far right of the bar, your account button signs you in and out. The start page has:
- New report opens the New report dialog over the page, the same one as a workspace's Reports page (see Make a report).
- Open… opens an
.ifprfile, or an older report package (.package.json) exported from the website. It also converts an SSRS report (.rdl) and a Power BI template (.pbit) or project (.pbip) (see Converting an SSRS report and Converting a Power BI report). - Recent lists the files you opened lately, one per line with its folder beside it. Click a file's name to open it. Remove takes one off the list; it doesn't delete the file.
- Your workspace (when you're signed out) offers Sign in and See the plans.
- Your workspaces (when you're signed in) lists every workspace you belong to. Open opens its reports, datasets, Query Workbench, data sources and Ask, as on the website. A Plus or Ultimate workspace's library on this computer is listed in the same pages, beside what's online.
A report has the window to itself: the sidebar and top bar give way to a strip with ‹ Home, the file's name and your account button. To get back to the start page without closing the report, choose ‹ Home: the start page opens with the sidebar, and the sidebar's Desktop section gains an item named after your report, between Home and New report, that takes you back to it. From there, New report, Open… and Recent replace the report in this window. If it has unsaved changes, the app first asks whether to Save, Don't save or Cancel. It asks the same before you leave the report for a library, Settings or Sign in.
A file you double-click while the app is running opens in this window when it has no report, and in a new window when it has one.
Make a report#
- Choose New report, on the start page or in the sidebar. The New report dialog opens over the page you're on. It is the same dialog as on a workspace's Reports page.
- Type a Report name, and choose the kind: Dashboard, or Printable report (a paginated document: invoices, statements, lists). Both can be made signed in or not.
- Start from begins at Blank report: one empty page, ready for data from Add data. For a dashboard it also offers With data, Spreadsheets and From a Power BI report (.pbit or .pbip) (see below); for a printable report, a table of the first dataset, the templates that fit your data, and From an SSRS report (.rdl) (see below). A dashboard has Pages for new datasets too: A page for each dataset or No pages.
- Once Start from takes data, tick where it comes from. Each place shows only when you have something there:
- Datasets: your workspace's datasets, signed in. The live workspace's and your library's are listed side by side ("Workspace: the live workspace", "Workspace: its library on this computer"); pick from either or both, and each picked one is marked Live workspace or Library. The report reads them each time the file opens. A live one can instead be Read once and kept in the file: its rows are saved in the file, and Refresh reads them again.
- Linked data: signed in, once your workspace has a data source. Pick the live datasets a data source fills, or New from a data source… to make one: it is saved in the live workspace straight away.
- Local database: pick a saved connection (this computer's, or a library's) and the database dialog opens on it, or choose New connection… (see below). Each read is listed.
- Spreadsheet: drop Excel or CSV files on the box, or click it to choose them. Each file is listed with its tabs under it. The first file's name becomes the report's name until you type one. What you pick is listed under its source with a trash can (Remove from this report), which only takes it out of this report: nothing is deleted from a workspace, a library or a connection. On a dashboard each one has Create a page for this dataset, which starts at Pages for new datasets.
- On a dashboard with a spreadsheet or a database read, Data to keep: Keep the rows in the file (the default), or Keep only the columns: the file stays small, asks for each spreadsheet again when it opens, and reads each database again.
- Choose Create report. To open a report you already have instead, choose Open a report… at the bottom left of the dialog. Cancel closes the dialog and leaves the page and any open report as they were.
- Build the report as you would on the website. The builder is the same: see The report builder.
- Choose Save. A new report is always a file on this computer until you save it. The first time, Save the report asks where it goes (see Where a report is saved): A file on this computer (the default: then pick a folder and a name; the file ends in
.ifpr), The library or The live workspace.
A file keeps the rows of the library and live workspace datasets it uses as they were last read, so it opens offline. When it opens, it reads each one again with your own account, and from then on the dataset works as it does in a library or live report: you can edit its rows and run its write actions (saved to the live workspace or the library, then read again, and the file is marked changed so Save keeps the new rows), change its settings, use Edit source… and Set up write-back… where you may change the dataset, and it updates when someone else changes it. A dataset the file can't reach when it opens (you're offline, signed out, or no longer have access) shows the rows the file stored, read-only, and the file says when they were read and why, for example "Orders: showing the rows stored on 1 Nov 2026, 10:00; sign in to read it". Refresh dataset in Add data ▾ reads it again once you can reach it. A new file whose dataset can't be read opens with its columns and no rows.
A report has the report builder's toolbar (see The report builder), with three of the desktop's own controls added. The file's name, with a • while there are unsaved changes, is in the window's title.
| Where | Control | What it does |
|---|---|---|
| Centre | Add data | From a spreadsheet… (in a printable report, held in memory while it is open) and From a database… |
| Centre | Refresh data | Beside Add data, shown when the report has a dataset that can be read again: Refresh dataset for each one (in a printable report, only the ones whose rows it keeps: the rest are read each time it runs), with Reload all of dataset for one that reloads part of its data |
| Right | Save | Saves the report where it is (Ctrl+S): its file, or the library or live workspace it was saved to. The first time, it asks where it goes |
| Right | Save as… | Asks where again and saves a copy there; the window then works on the copy |
Where a report is saved#
The first Save of a new report, and every Save as…, open Save the report with three places. One you can't use is greyed, with the reason (you're signed out, there's no library on this computer, your workspace's plan keeps no reports, or your role only lets you view).
- A file on this computer (the default): pick a folder and a name. Saving again writes the file without asking.
- The library: pick the library (when you have more than one) and a Folder you can edit. The dialog lists each dataset with what it becomes in the library: rows kept in the file are stored in the library; a spreadsheet kept as columns only becomes a library dataset with its columns only; a database read each time is read through the library's connection (a query you typed is saved as a query text there, which each computer approves once); a dataset of the library stays as it is. A live workspace dataset stays in the live workspace: the library report reads it from there with each colleague's own account, so colleagues who can't reach that workspace's datasets see its columns with no rows, and the report says it "isn't accessible in your current workspace". A dataset of another library goes in as a copy of the rows it shows. Choose Save.
- The live workspace: the same steps as Publish, titled Save to the live workspace: the workspace, the folder, and what each dataset reads there.
Once a report is in the library or the live workspace, Save saves it there, over the same report. Its datasets stay as they were saved: if you add a dataset after saving to the live workspace, Save shows the live workspace's steps again (the report there is replaced); after saving to a library, use Save as… to save it there again with the new dataset.
A file opened without some of its datasets (see the five-dataset limit above) has no file yet: its first Save asks the same, and a file starts in the original's folder under the new name.
From Home, the sidebar does the rest. While a report is open and you're signed in, its Desktop section also has Publish… (when you have a Mini, Plus or Ultimate workspace to publish to) and Add to library (Plus and Ultimate).
The Data pane's Add dataset offers the same data, and more: From a spreadsheet, From a database and, signed in with a library, From a library, a dataset kept in your library, read from there whenever the report opens. Under each dataset's name, the Data pane says where its rows come from: kept in this report, read from a database or a library, or read when the report runs.
Signed out, a report on this computer can use up to 5 datasets: New report and Add dataset stop there and say why. Signed in, your plan's limit applies (none on Plus and Ultimate). Opening a file with more than 5 while signed out asks first: "This report has n datasets. Signed out, a report can use 5." Tick the ones to remove (each says how many visuals use it; those visuals stop working, and show that their dataset is missing) and choose Open: the report opens without them as Name (5 datasets), with no file, so your first save asks where (it offers Name (5 datasets).ifpr beside the original). The original file isn't changed unless you pick it yourself in that save and replace it. Sign in opens the whole file once your plan allows it; Cancel opens nothing.
Until you sign in, a report's page templates are the two the free trial has (Explore and Category Breakdown). Signed in, you get your plan's templates.
Printable reports on this computer#
A printable report opens in the printable report designer instead of the report builder, with the same ‹ Home strip, Add data in the middle of its toolbar and Save and Save as… on the right. Design it, preview it, and save it as a PDF or export it to Excel exactly as on the website; no account is needed. An .ifpr that holds a printable report opens in the designer the same way.
A printable report keeps no rows in its file: every dataset is read when the report runs. A database read is read again each time you run the report (and when a parameter changes); a library or live workspace dataset is read from there. When you're signed in to a Plus or Ultimate workspace, which has a library, a spreadsheet has two choices in Add dataset, From a spreadsheet. Signed out, or on Free or Mini, there's no library, so the spreadsheet is held in memory and the choice isn't shown:
- Hold in memory while the report is open: each tab becomes a dataset, and its rows stay in the window. The file keeps the columns and never the rows, so the rows are lost when the report closes. Dropping a spreadsheet on a printable report's window, or Add data ▾, From a spreadsheet…, does the same.
- Store it in a library: each tab becomes a dataset in the library, and the report reads it from there when it runs.
When you open a report that holds a spreadsheet in memory, the dataset has its columns and no rows. The Data pane says Pick Sales.xlsx again under the dataset, and the preview says the same above the sheets, beside the parameters. Click it and choose the file: its tab with the dataset's name (or the first tab with all the dataset's columns) fills the dataset. A file that lacks one of the columns is refused, naming the missing columns. Nothing shows from that dataset until you pick it, and a run without it is noted in the file's run log as a failed read ("no rows: the spreadsheet isn't loaded"). Picking the file again isn't a change to save.
Publish… copies the rows a held spreadsheet has in memory. If you haven't picked it again yet, its row in the list says Pick the spreadsheet first with the same button, and Publish waits. To a Mini workspace it goes up with its columns only, and the web report holds the spreadsheet in memory the same way.
What needs a workspace is left out: Schedule (sending the report by email) and Ask FastPass. Publish the report to a Plus or Ultimate workspace to use those.
Pictures a report takes from web addresses are fetched by the app itself while you're signed out, so a site that doesn't allow it shows a crossed box in their place; signed in, they come through Insight FastPass.
Converting an SSRS report#
SQL Server Reporting Services reports (.rdl) convert to printable reports on this computer, without an account:
- Open… an
.rdlfrom the start page or the sidebar, or drop it on the window. Its subreports, shared datasets (.rsd) and shared data sources (.rds) are read too when they are in the same folder (or dropped with it). - Or in New report, choose Printable report, then Start from From an SSRS report (.rdl), and drop the file on the box under it, or click the box to browse (with its subreports and shared files, or a
.zipof them). - Or just drop the
.rdlon New report's spreadsheet drop area (Start from Spreadsheets): the dialog switches to Printable report started from the SSRS report by itself.
The report opens in the designer, and a Converted from SSRS window lists:
- Datasets to fill: the report came with each dataset's columns, not its rows. Choose From a database… beside each one (read again each time the report runs), or From a spreadsheet, into the library… (signed in with a library) or From a spreadsheet, held in memory… (without one); columns with the same names bind as they did in SSRS. The dataset's SQL is in the notes, to write the database read from.
- Conversion notes: what to set up, what was left out and what was approximated, each with where it was in the SSRS report and the original expression. Download notes saves them as a Markdown file.
Conversion notes in the strip at the top opens the window again, and says how many datasets are still without data.
Converting a Power BI report#
Power BI templates (.pbit) and projects (.pbip) convert to dashboards on this computer, without an account. A .pbix can't be converted: save it as a template in Power BI Desktop first (see Save your report as a template).
- Open… a
.pbitor a.pbipfrom the start page or the sidebar, or drop it on the window. A.pbipis read with theName.ReportandName.SemanticModelfolders beside it, straight from disk, so there's nothing to zip. - Or in New report, choose Dashboard, then Start from From a Power BI report (.pbit or .pbip), and drop the file on the box under it, or click the box to browse.
- Or just drop the
.pbiton New report's spreadsheet drop area (Start from Spreadsheets): the dialog switches to the Power BI report by itself.
The dashboard opens in the builder, and a Converted from Power BI window lists:
- Datasets to fill: the report came with each table's columns, not its rows. Choose From a database… beside each one: when Power BI read a SQL Server table, the database dialog opens with the server, database and table filled in, so you pick a connection and import. Or fill it from a spreadsheet: each dataset says which file and sheet Power BI read.
- Conversion notes: what to set up, what was left out and what was approximated, as on the website. Download notes saves them as a Markdown file.
What comes across, and what doesn't, is in Import from Power BI.
If the app closes before you save#
The app keeps a copy of each unsaved report every minute. If it closes before you save, the start page says "A report was not saved when the app last closed" the next time. Open it brings the report back with its changes; Discard throws the copy away.
Read a database#
The app connects from your computer, so it reaches any database your computer can: a server on the office network, through your VPN, or on your own machine.
- In a report window, choose Add data ▾, From a database… (or Connect… on a new report). Read from a database opens.
- Under Connection, choose New connection… the first time.
- Fill in the connection (see Connections below), choose Test, then Save.
- Choose the table, view or procedure, either way:
- Type its name under Table, view or procedure name and choose Check (or press Enter). You don't have to wait for the list, which can take a while on a large server. Type the name without
EXEC; the schema can be left out when it's the usual one (dboon SQL Server,publicon PostgreSQL, your own on Oracle). Check finds it, then reads its first rows, so a name that doesn't exist, or that your account can't read, says so straight away. - Pick it from the list under Or pick from the list, which loads in the background while the dialog is open. Search to narrow it.
- Or a saved query text: a statement saved on the connection and approved on this computer (see Query texts). Its parameters show under Parameters: type a value for each (several values separated by commas for a list; a date also takes a relative date in a filter). Choose Preview to see its columns, then tick the ones you want. Query texts can be read on SQL Server, PostgreSQL and MySQL connections.
- Or type a query: opens a box for your own SQL. It must be one read: a
SELECT(aWITHtoo), or on SQL Server anEXECof a procedure; anything that writes is refused, with the reason. Each@Namein it shows under Parameters with a type and a value to give it. Choose Read: the app checks the query, then reads its first rows, and its columns show under Columns of. If you change the query, choose Read again. There's no approval step. In a report file the query is kept in the file; in a library it's saved as a query text on the library's connection, so everyone using the library sees it. Anyone else who opens the report (or you, on another computer) is asked once before it runs there: "This report runs this query against server · database: … Run it?". Run it reads it; Don't run it opens the dataset with its columns and no rows, and the report says so. Choosing Refresh asks again.
On SQL Server, a stored procedure can be read too. What Parameters shows depends on the report.
In a dashboard, you give the procedure its values once, here: a box for each parameter, with a NULL tick box beside it. A parameter left empty uses the procedure's default; one with no default needs a value or NULL, and Import waits until it has one ("@Name has no default: give it a value, or pick NULL."). These are the only values the dashboard ever reads with: no filter, slicer or reader changes them, and Refresh runs the same read again. The Data pane says what it reads, for example "Runs dbo.SalesByRegion with @Region = West, @Year = 2026". Preview reads the first rows with the values given (NULL for an empty one with no default).
In a printable report, Parameters lists the procedure's parameters, as SSRS does: you never type their values here. Each one becomes a report parameter named after the dataset and the procedure's parameter (
Sales_Regionfor@Regionon the dataset Sales). The dataset is added with its columns, and nothing is read until the report runs; then it reads the procedure with the values picked when you choose Run (at the right of the parameter bar), and again each time you change one and choose Run. A parameter left blank uses the procedure's default (@Region nvarchar(20) = NULL), and NULL can be picked for any of them. You can delete any of these parameters, even one the procedure needs: the report then sends NULL, and if the procedure needs a value the read fails and says why. The Parameters tab lists inputs no parameter sends a value to, with Add a parameter for it. Preview reads the first rows with each parameter's default, NULL where it has none. - Type its name under Table, view or procedure name and choose Check (or press Enter). You don't have to wait for the list, which can take a while on a large server. Type the name without
- Tick the columns you want under Columns of table. All and None tick or clear every one. Columns of a type the app can't load (images, binary data) show "(not loaded)" and can't be ticked.
- To read only some rows, use Only rows where and Add a filter: pick a column and a condition such as equals, contains or between. A date column also takes relative dates such as today or a number of days ago, worked out each time the dataset is refreshed.
- Name the dataset under Dataset name.
- In a dashboard (or a library), Later refreshes says how each refresh after this read lands: Replace every row, or Reload part of the data (see Reload part of a dataset). This read is always the whole first load.
- Choose Preview to see the first rows as they will be stored, then Import.
The dialog stays open while it reads, and counts rows as they arrive. If the read fails (a wrong value, say, or a table your account can't read), the reason shows at the top of the dialog with everything you chose still filled in, so fix it and choose Import again.
In a printable report, the procedure's parameters are on the designer's Parameters tab like any other: give each a label, its Available values (typed in, A list I type, or From a query: a dataset's value and label fields), a default, and whether it is Required, allows a null or blank value, and is Visible above the report. A null or blank value that's allowed is always offered as (Null) or (Blank), even when the list doesn't include it; null reaches the procedure as NULL. A read stops after 10 million rows, 4 GB or an hour. If it reaches one of these, nothing is imported, so add filters or fewer columns and try again.
You don't write SQL to read a table: you choose an object, its columns and filters, and the app builds the read. Tables and views (and SQL Server stored procedures) can be read this way, as with the website's data sources. When a read needs more than that, type it with Or type a query, or whoever manages the connection can save a query text on it.
Edit a dataset's source#
Edit source… (in the Data pane under a dataset read from a database, and beside each dataset on a library's Connections page) opens Read from a database filled in with what the dataset reads: its connection, table, view, procedure or query text, columns, filters and values. Change any of them; for a library dataset, you can also choose whether its rows are kept in the library or read when its reports run. Save the new source first names any column the report (on the Connections page, any report in the library) uses that the new read doesn't have, and asks before going on. A dashboard's dataset is read again at once and changes only once that read has finished: if it fails, the dialog says why and the dataset keeps its old rows and source. A printable report's dataset takes its new columns, and the next run reads it; its report parameters follow the new source's parameters.
When you edit the source of a dataset that reloads part of its data, Later refreshes keeps its rule. If the new source no longer has a column the rule reads, the dialog names it ("The reload rule reads OrderDate, which the new source doesn't have…") and the dataset goes back to replacing every row until you set the rule again. Saving the new source reads every row once either way.
Refresh a dataset#
Datasets read from a database remember where they came from, and so do copies of live workspace datasets and datasets from a library. In the report window, Refresh data ▾ (beside Add data) lists Refresh dataset for each one. A printable report lists only the ones whose rows it keeps, such as a library dataset's: a database read when the report runs has no Refresh, since every run reads it again. Refreshing reads the table again with the same columns and filters and replaces the dataset's rows, or, for a dataset that reloads part of its data, only the rows its rule covers. If you've edited its rows in the grid, the app asks first, and says which rows lose their edits: every row on a whole refresh, or only the rows the rule replaces (edits to other rows are kept). The new rows replace the old ones only once the read is whole: a read that fails leaves the rows you had, and says why. The report keeps its pages, visuals and settings. Each refresh and each read when the report runs is noted in the file's run log, with how it landed and how many rows it removed, and the Data pane shows the latest.
A dataset the report reads from a library dataset that refreshes on a schedule shows, in the Data pane, that dataset's schedule and last run ("Refreshed daily at 06:00; last run 31 Oct, 06:00: 1,310 rows"). Refresh now appears only when it is behind: its time has passed with no refresh running, or its last refresh failed. It runs the library dataset's own refresh, then the report reads the new rows.
Reload part of a dataset#
A large dataset doesn't have to be read whole on every refresh. Choose Reload part of the data (under Later refreshes when you read a database, as Rows when you read into an existing library dataset, or under Each refresh on the Connections page), then tick one or both:
- A date window: pick the date field and the window (Today, The last few days, This week, This month, This quarter, This year, Last month, Since a date, Between two dates). Each refresh reads the rows dated in the window and replaces the stored rows in it, whatever they were, so a day with no rows at the database ends up empty. "The last 5 days" means five days including today; "between 1 Oct and 1 Nov" runs up to but not including 1 Nov. For a procedure or query text, Give the bounds to parameters passes the window's start and end to its parameters, so it reads only the window. Remove rows older than also drops old rows on each refresh.
- Matching keys: pick the columns that identify a row (one to eight). Every stored row whose key comes in again is replaced; the rest stay. Rows deleted at the database stay until you reload everything.
- Rows marked as deleted: a column that marks a deletion ("when it is true", or "when it has a value"). A marked row removes its match and isn't stored.
A filter or parameter can also use since the last run: the moment the last refresh started reading, less ten minutes, so ModifiedAt on or after the last run reads just what changed.
The first refresh of a new dataset reads every row; later ones read only what the rule says. Reload all of dataset (in a report window's Refresh data ▾) and Reload everything (on the Connections page and in the Data pane) read every row once and replace the dataset whole; the next refresh follows the rule again. A refresh whose rows don't fit the rule (a row dated outside the window, or with no key) changes nothing and says which row. A report file keeps each dataset's last refresh and the moment "since the last run" reads from, so it reloads correctly when opened again on any computer.
The .ifpr keeps the connection and the read, but never the password. On another computer, a refresh that needs a password can't sign in until one is entered there: choose Add data ▾, From a database…, pick the connection under Connection, choose Edit, enter the password and Save, then refresh. A connection that uses My Windows sign-in signs in as whoever opened the file.
Connections#
A connection you add is saved, so you only enter it once. One made for a report on this computer is kept by the app and offered for every report here (and the report's .ifpr keeps a copy, without the password). One made in a library is kept in the library and shared with everyone who opens it. You can add to a library's connections from its live side too, once the library is on this computer. The password stays in this computer's keychain either way.
| Field | What to enter |
|---|---|
| Database | SQL Server, PostgreSQL, MySQL / MariaDB, Oracle or ODBC |
| Name | A name for you, such as "ERP reporting" |
| Server and Port | The server's name or address, and its port. The default port fills in for each database |
| Database (Service name for Oracle) | The database to read. After a successful Test, the databases you can see are offered as buttons |
| Data source name (DSN) or Or a connection string | ODBC only: one or the other. Leave the password out of the connection string |
| Sign in with | A user name and password; My Windows sign-in (SQL Server on Windows: no password at all); What the data source has (ODBC, when the DSN holds the sign-in) |
| User name and Password | The database account. A password you save is kept in your computer's own password store (Windows Credential Manager, the macOS Keychain or the Linux secret service), never in a file |
| Encryption | Connections are always encrypted. The server's certificate is checked against the authorities your computer trusts, or a CA you supply (paste its PEM certificate). Trust server certificate skips that check (see below) |
Test signs in and reports "Connected as account" with how many tables and views that account can read.
A server with its own certificate#
Many servers inside a company use a certificate your computer doesn't trust, often one the server made itself. Test then shows "The server sent a certificate this computer does not trust", with who it is issued to and by, its names, dates and SHA-256 fingerprint. Check the fingerprint with whoever runs the server, then choose Trust this certificate. From then on the app trusts that exact certificate for this connection, and nothing else. Stop trusting it undoes that. When the server's certificate is renewed, you'll be asked again.
Trust server certificate#
Tick Trust server certificate (SQL Server, PostgreSQL and MySQL) to accept whatever certificate the server shows, as the same option in SQL Server Management Studio does. The connection is still encrypted, but the app no longer checks that it's talking to the real server, so someone who could redirect your network traffic could pretend to be it. Use it on a network you trust. Trusting the certificate a Test shows (above) is safer, because the app then accepts only that one certificate.
Oracle doesn't offer it: its client always checks the certificate, so trust the certificate a Test shows instead. For ODBC, the driver's own settings decide (for example TrustServerCertificate=yes in a SQL Server connection string).
Oracle connections use TCPS (encrypted) on port 2484 by default.
Which addresses it will connect to#
The app reads databases on your network and the internet, but refuses cloud metadata and other reserved addresses, so a connection can't be pointed at your computer's cloud credentials.
Query texts#
A query text is a statement you type and save on a connection, with @Name parameters for its values: a read (a SELECT) to import rows, or a write (an INSERT, UPDATE, DELETE or MERGE, or on SQL Server an EXEC of a procedure) for a write action. Values always go to the database as parameters and are never written into the text.
To add one, edit a saved connection: under Write-back on this computer, Query texts, choose New query text. Give it a Name, choose what It does, and type the Statement. Find the text's parameters adds a row for each @Name it uses; for each, choose its type, whether it takes a List (and its Delimiter, except on PostgreSQL, which takes a list as an array: = ANY(@Ids)), whether it is Required, and a Default. Then choose Check and save.
The app checks the text before it saves it:
- It must be exactly one statement of the kind you chose. A second statement, a change to the database's structure, a transaction,
USE, dynamic SQL and the like can't be saved, and the reason shows. - It is scanned for patterns that are often a mistake or an attack: a statement hidden after a semicolon or in a comment, a condition that is always true (
OR 1=1), aUNION SELECTreading the system catalog, and so on. Some patterns are never allowed (reading files, delays, another database, text spelled withCHAR()calls); the text can't be saved until you rewrite it. Others are worth a second look: the text is saved but Waits for approval, with each finding listed by line. Choose Approve on this computer once you've checked each one is meant.WHERE 1=1 AND …, for example, is a common, harmless habit.
A text runs with everything the connection's account may do, so an account that can only read is the protection for reads. Editing a text needs approving it again. Revoke stops it running here; Delete removes it.
Save edits back to a database#
A dataset read from a database can carry write actions: what a grid edit, a new row or a deleted row does in the database (an update of a table, a stored procedure, or a query text). They're set up in the dataset's Write-back section, as on the website; on this computer they write through one of your connections. When you choose Save changes, the app runs the dataset's actions in order, in one transaction per connection, then saves the edits to the dataset itself.
Approval on this computer. An action runs through a connection on this computer only once it is approved here, for exactly what it does. When you save a new or changed action, the app shows what it runs (the statement and what fills each value) and asks Approve a write action. If you decline, or someone else's action reaches this computer, a save stops at that action with "isn't approved on this computer": the failure message offers Review and approve on this computer, and the connection lists it under Write actions waiting for approval. A changed action needs approving again. Revoke under Approved write actions stops it. On a library's connection, approvals are kept with the library, so they apply to everyone who opens it.
The first save is a dry run when the action has Dry run next save ticked: the database checks every change and then undoes them, nothing is saved, and your edits stay in the grid with a notice. The dry run is then turned off: choose Save changes again to write them.
If a step fails, a message says which action and why; steps already done stay done, and your edits stay so you can fix them and save again. A row that no longer exists in the database is reported as not found, and the rest carry on. Every row each save changed is kept in the dataset's Write-back history, and an .ifpr keeps it too.
Sign in#
Signing in is optional. Choose Sign in on the start page, in Settings, or from your account button in a report window:
- Your web browser opens Insight FastPass's Sign in to the desktop app page. If you aren't signed in to the website in that browser, sign in first, as usual.
- Check the account it names and choose Continue. (Use a different account signs that browser out so you can sign in as someone else.)
- The browser says the app is signed in. Close the tab and go back to the app: every open window has picked up the sign-in, and the report you were working on is still open.
The page only works for the app that asked: it hands a one-time code to the app on this computer, which expires after two minutes.
If you'd rather not use the browser, choose Sign in here instead. A small sign-in window opens inside the app; sign in as you do on the website, and it closes by itself when you're done.
Sign out (your account button, or Settings) signs out in every window, also without closing your report. Reports saved as files keep working.
Signed in, you get:
- your plan's page templates in reports saved as files;
- a library for each Plus or Ultimate workspace you belong to (see below);
- Publish…, to copy a report and its data to a workspace.
Settings, Account shows who you're signed in as, Manage your workspace on the web and Sign out.
Working offline#
The app remembers your sign-in for 30 days without reaching the internet. In the last seven days the start page reminds you to sign in again. After 30 days your libraries close until you sign in again. Reports saved as files always open, signed in or not, so save a library report with Save as .ifpr before you go offline for longer.
Your library#
A library is your workspace's datasets and reports kept on your computer instead of in the cloud. It works through the same screens as the website: Datasets, Reports, the Workbench, Calculated Fields and Summary Calculations, dataset connections and folders. A library has no limit on how many datasets or reports it holds, and every dataset keeps its rows, whatever the plan's limits online. The plan's features still apply: snapshot datasets and saved versions need Ultimate, row access Plus or Ultimate.
Plan availability One library for each Plus or Ultimate workspace you're a member of, signed in. Viewers can read a library; editors and above can change it.
Open it: Live and This computer in one place#
On the start page, under Your workspaces, choose Open (or pick the workspace from your account button in a report window). The workspace's pages show what's online and what's in its library together:
- Reports has two roots in its folder tree: Live (the workspace online, as the website shows it) and This computer (the library). Each has its own folders; moving a report or folder stays within its root, and a library report goes online with Publish….
- Datasets lists both, each marked Live or This computer. The buttons above the list show All, Live or This computer, and the app remembers your choice. Put new datasets in says where an upload or a database read goes; it starts from the last place you chose.
- The Query Workbench has a Live / This computer switch at the top: queries run against the one you pick.
The bar at the top has:
- Connections, which lists the library's database connections.
- Publish… and Save as .ifpr while a library report is open, and Save a local copy… while a live one is.
- Your account button (far right): who you're signed in as, your workspaces, Settings, and Sign out.
Sharing, public links, favorites, scheduled deliveries and saved versions are for reports in the live workspace. A library report doesn't offer them: publish it to use them.
The sidebar on the left starts with the same Desktop section as everywhere in the app: Home (the start page), New report, Open… (an .ifpr) and Settings. Support ↗, Docs ↗ and Community ↗ sit at the bottom and open the website in your browser. Your account button is at the far right of the bar, after the tutorials' ?.
Signed in, the sidebar's Reporting section is as on the website: Reports, Datasets, Data sources, Query Workbench, Ask and, once Ask FastPass is turned on for the workspace, Ask history. Data sources, Ask and Ask history work on the live workspace. When you add a data source in the app, its last step (entering the database password, which goes straight to the sealed credential service) is Finish on the website ↗: the data source is saved, and the website opens at that step.
Running the workspace (people, invitations, security, plan and billing, your settings) stays on the website: use Your account ↗ from your account button.
Read a database into a workspace#
In the app, Datasets has From a database… beside the file upload, and so does a workspace report's Add dataset dialog on the Data pane (the report builder's and the printable report designer's), where the new dataset is added to the open report. A stored procedure with parameters works as in a report on this computer when the new dataset goes to This computer. Added to a dashboard, it takes its values in the dialog, as above, and the dashboard always reads with them. Added to a printable report, the dataset is added with its columns and no rows, the report gets a parameter for each of the procedure's parameters, and each report reads the procedure with its own values when it runs, so one report's values never change another's rows. From Connections or Datasets (no report yet), a value you give is always used and one left empty is left to the reports: a printable report asks for it when it runs, and a dashboard reads the procedure's default, NULL where it has none. Connections lists a dataset read when its reports run with no Refresh. On Live, the website can't reach a database on your network, so the procedure is read once, with the values you give (empty: its default, NULL where it has none), and copied up; to pick its parameters each time a printable report runs there, use a live dataset from Data sources. It opens the same dialog as in a report window. With Put new datasets in on This computer, the new dataset goes into the library. On Live, the app reads the database on your computer and copies the rows up as a new dataset of the live workspace, counting toward its plan; to bring newer rows in later, read it again. New report reads a database into the new file instead (its Local database source), and Open a report… at the bottom left of the dialog opens an .ifpr in the report window.
Read a database into the library from Connections#
- In the workspace's pages, choose Connections in the bar at the top.
- Choose New connection to add one, or Read from a database… to read a table. The dialog is the one described in Read a database.
- Under Into, choose A new dataset and name it, or pick an existing dataset. For an existing one, Rows is Replace every row, Add as the next snapshot for a snapshot dataset (Ultimate), or Reload part of the data. A new dataset can set Later refreshes at once.
Datasets read from a database lists each one with its table, connection, row count and when it was read, its last refresh and when the next is due. Refresh reads it again by its rule; a read that fails leaves the rows it had. Under it:
- Each refresh: Replace every row, Add as the next snapshot (a snapshot dataset), or Reload part of the data, with the rule under it (see Reload part of a dataset) and Save how it refreshes. Saving a date window offers to add an index on its field, so a reload never scans the whole dataset.
- Schedule: Only when asked, Every … (15 minutes to a day), Daily at a time, or Weekly on some days at a time, in this computer's time zone. "Runs while the app is open or in the tray", with a link to Settings: closing the app's last window leaves it in the system tray while a library has a schedule (see Scheduled refreshes). A refresh missed while the app was closed or the computer asleep runs once when it starts or wakes, then the schedule carries on from then. The first time you set a schedule, the app asks whether to start when you sign in.
- When a report opens: a report on it shows the rows it has, then the new ones once they're read.
- Allow edits: rows read from a database are read-only in the grid, since a refresh replaces them. Tick this to edit them anyway, knowing the next refresh replaces your edits (a reload replaces only the rows its rule covers, so edits to other rows are kept); or set up write-back to save edits to the database. A report's Data pane offers the same as Allow edits….
- Reload everything (for a dataset that reloads part of its data): reads every row once and replaces the dataset whole.
Each refresh, by hand or on its schedule, goes into the library's run log with how it landed and how many rows it removed, and a report's Data pane shows a dataset's latest, with why it failed when it did.
Scheduled refreshes#
A library dataset with a Schedule is refreshed by the app while it is open, and after you close its last window too: the app stays in the system tray (the notification area on Windows, the menu bar on macOS, the tray on Linux) while any library has a scheduled refresh. Its tray icon has Open Insight FastPass, Refresh now (every scheduled dataset, now) and Quit. Settings, Scheduled refreshes has two switches:
- Keep running for scheduled refreshes (on): turn it off and closing the last window quits the app, so schedules run only while a window is open.
- Start when I sign in (off): starts the app in the tray when you sign in to the computer, so the schedules run after a restart.
A refresh needs the connection's password on the computer that runs it (or a connection that uses My Windows sign-in). A scheduled refresh no one present can run is noted in the run log once a day: "Not refreshed on its schedule: no credential on this computer". On a shared library, one computer runs each scheduled refresh; see Share a library on your network.
When you read a database from a library report's Add dataset, From a database, choose how: Keep the rows in the library (refreshed as above), or Read them when the report runs (nothing is kept, each run reads the database).
A library's connections are shared with everyone who opens the library. Each person keeps their own password on their own computer.
Move a library#
A library starts in your computer's app data folder. In Settings, Libraries, choose Move… and pick an empty folder. The app copies the library there and works from the new place. When the library was in the app's own folder, a banner offers Delete it for the old copy. A library moved from anywhere else is left where it was, since others may still use it.
To share a library with your team on a network folder, see Share a library on your network.
Publish to your workspace#
Publish… puts a report in a workspace online, so people can open it on the website, and says for each of its datasets what it reads there.
- In a report window, or with a library report open, choose Publish….
- Pick the Workspace and Folder, and choose Next.
- The dialog lists every dataset at once, with what it reads ("reads dbo.Sales on ERP"). For each one, choose what it reads in the workspace. Each dataset offers only the choices that suit it, and starts at the first that needs nothing else picked:
- Upload with its rows, as a new dataset (for a database read each time, A one-time copy of its rows), or replacing a dataset of the same name you can edit.
- Upload columns only: a dataset with its columns and no rows, to import its data on the website. The only upload for a spreadsheet you kept as columns only.
- A linked dataset in its place: a dataset in the workspace that one of its data sources fills, read from your database through the gateway. One it was published to before, or with the same name, is picked already. The same table or procedure must exist on that side, or the report's first run there says it doesn't.
- Another dataset in the workspace, as it is.
- Send the query for approval, on data source: only for a query you typed on a local database. See below.
- A dataset that's already in this workspace stays as it is.
- A copy of a dataset read when the report runs (a printable report's, or a procedure) is read now: type the values that pick its rows (leave one empty for its default, or type
{null}), and choose Count rows to see how many go up. Its report parameters are taken out, since the copy's rows are fixed. - If the folder has a report with the same name, choose Replace it or Publish as Name (1). Then choose Publish.
What to know:
- A Mini workspace keeps no rows, so the report goes up with each dataset's columns only, and you import its data on the website. Free workspaces have nothing to publish to, so Publish isn't offered.
- The published report and its copies count toward the workspace's plan. Reading a dataset that's already there doesn't add one.
- Publishing again picks what each dataset went to before.
- Send the query for approval writes the query you typed as a new query text on the data source you pick (one you manage, for the same kind of database), which tests it, and creates the dataset in the workspace with the columns your read gave, reading that query text. The dataset is empty until an owner or admin of the workspace approves the query text on the data source's Query texts page; if you're an owner or admin, you can approve it yourself. The publish result says so, and so does the dataset in the report's Data pane until it's approved. Nothing reads it the moment it's approved: a live dataset fills the next time the report reads it (Read again, or opening the report), and a stored one when you choose Refresh now. A query with filters, or one whose values the report gives when it runs on a data source without live reads, can't be sent: upload a copy instead.
- The published report is separate from your file or library. Later changes on either side don't carry over until you publish again.
Keep a live workspace report on this computer#
With a live workspace report open, Save a local copy… saves it as an .ifpr:
- Keep reading the live workspace: the file says where each dataset is, keeps its rows as last read, and reads it through your sign-in when it opens. While you're signed in, its lightning mark in the Data pane, Read again, the "as of" time in the toolbar and Refresh live data while open work as on the website, and so do editing its rows, its settings and Edit source…. Signed out, or when the workspace can't be reached, the dataset shows the rows the file stored with the warning, and isn't read again until you refresh it.
- Keep a copy of the rows: the file opens anywhere, signed in or not, and Refresh reads them again from the workspace. A printable report keeps no rows either way.
Ask FastPass about a report on your computer#
Signed in to an Ultimate workspace where Ask FastPass is on, a report window and a library report have Explain this page, Ask about this report and Prompt a report page in the strip above the report. They open the Ask panel over the report, so the report stays open.
- Explain this page sends the values the open page shows, and the answer appears in the panel.
- Ask about this report takes your question. Choose Text answer for a written answer, Page in this report for a new page, or Calculations for Summary Calculations and calculated fields. A query isn't asked from a report: ask for one on the Ask page of a live workspace, or with Ask FastPass in the Query Workbench.
- Prompt a report page asks what the page should show. FastPass plans it from the report's data.
For a page, the panel lists the visuals FastPass planned. Add page puts the page after the page you were on. Undo takes it out again, and the page is saved with the report like any other.
For calculations, the panel lists each one FastPass proposed, checked against this report, with the sample FastPass got when it ran the formula on your data. One that can't be added says why (a name the report already uses, or a formula that doesn't check here). Untick any you don't want, then Add to this report:
- In a library report, Summary Calculations are stored in the library, so its other reports can import them, and calculated fields are saved on their dataset when you may change it.
- In a report window (an
.ifprfile), both are kept in the file and saved with it.
Undo takes the Summary Calculations out of the report again; remove a calculated field from the Data pane. Copy formula copies one to add by hand.
A query asked on a live workspace's Ask page in the app opens by itself in a Query Workbench window of its own as soon as it is ready, so the window you asked from stays as it was.
Your data stays on your computer. FastPass's questions about it (totals, breakdowns, column summaries) run here, on the copy the app keeps, and only their results go to FastPass. In a report window, the app first copies the report's data into its local store, which takes a moment for a large file. The question is asked through your workspace: it counts toward the workspace's Ask usage and is kept in its Ask history, where it shows the file or library it was about. Because the data isn't in the workspace, it can't be re-run or followed up from there. Ask again from the report instead.
Keep the app open until the answer comes: closing the panel is fine, but FastPass reads the data through the app while it works.
Open and save .ifpr files on the website#
The website's report builder reads and writes the same files. In the builder toolbar:
- Export, Download report (.ifpr) downloads the report with its data and edits (Plus and Ultimate). On the Reports page, a report's actions menu has Download (.ifpr) too, without opening it, and Upload report saves an
.ifpr, a definition, an SSRS report or a Power BI template as a new report in the workspace. - Export, Import… opens an
.ifpr, a definition or an older package file (Mini, Plus and Ultimate).
So you can build a report on the desktop, save it, and open it on the website, or the other way round. See Exporting and importing.
An .ifpr of a dashboard holds the rows its datasets keep in the report. Treat it as you would the original spreadsheet or database extract. A printable report's file, and a dataset read when the report runs, hold no rows: only where to read them from.
Print and export#
Print opens the builder's print preview, as on the website, then your system's own print dialog, where you can also save a PDF. Export data and the grid's Excel export ask where to save the file.
Zoom#
Make everything in the app bigger or smaller, as in a browser:
- Ctrl + scroll, or pinch on a trackpad.
- Ctrl + = to zoom in, Ctrl + - to zoom out, Ctrl + 0 for 100% (⌘ on a Mac).
- The zoom control in the window's bottom-right corner: the magnifier buttons zoom out and in, and the percentage between them goes back to 100%.
The zoom runs from 50% to 200% and is remembered the next time you open the app.
Updates#
Settings, About shows the version. Check for updates looks for a new one and Install and restart installs it. If your version is older than the website supports, the start page asks you to update; your files and libraries keep working meanwhile.
The offline beta#
The offline beta, Insight FastPass Beta, is a copy of the app for trying it out without an account. It works entirely on your computer and never connects to the internet.
- What you get. One workspace, My library, with everything a Plus or Ultimate library has and no limits. Datasets are shared between your reports, and you can read databases on your network, save edits back to them, and make printable reports. You can also save and open
.ifprfiles, read these guides from Docs in the sidebar, and move your library to a network folder (see Move a library), so a team running the beta can work in one library. - What it leaves out. Signing in, the live workspace, publishing, data sources and gateways, Ask FastPass, sharing and public links, favorites, scheduled delivery, and Support and Community. The guides still describe these, as parts of the full product.
- When it ends. The start page shows the date the beta runs until. After that day, My library closes, but its reports and data stay on your computer and your
.ifprfiles still open. Installing a newer beta opens the same library again.
The beta installs beside the full app and keeps its own settings and library, so the two never mix.
Plan limits summary#
| Feature | Not signed in | Free and Mini | Plus | Ultimate |
|---|---|---|---|---|
Reports saved as .ifpr files | Yes | Yes | Yes | Yes |
| Datasets in a report saved as a file | No limit | No limit | No limit | No limit |
| Read databases on your network | Yes | Yes | Yes | Yes |
Convert SSRS (.rdl) and Power BI (.pbit / .pbip) reports | Yes | Yes | Yes | Yes |
| Page templates in a report saved as a file | Explore and Category Breakdown | The plan's | The plan's | The plan's |
| A library on your computer or a network folder | No | No | Yes, no limits | Yes, no limits |
| Snapshot datasets and saved versions in a library | No | No | No | Yes |
| Row access in a library | No | No | Yes | Yes |
| Publish to a workspace | No | Free: no. Mini: the report with its datasets' columns | Yes, within the plan's limits | Yes, within the plan's limits |