Installing

PQDriverless

Not part of the Power Query standard library. This function ships in PQDriverless, an open-source package of pure M readers for binary file formats — no ODBC driver, no OLE DB provider, nothing to install. Paste the M source into a blank query, or load PQDriverless.mez as a custom connector. See Installing.

There are two ways to get a driverless connector into a query, and they suit different situations.

	Paste the source	Load PQDriverless.mez
Setup	none	copy one file, change one setting, restart
Scope	the one file you paste it into	every file on that machine
Hosts	anywhere M runs — Desktop, Excel, dataflows, the Service	Power BI Desktop only
Works in the Service	yes, the M travels with the query	only through a gateway configured to allow custom connectors
Works on a locked-down desktop	yes	needs the Data Extensions setting changed
Updating	re-paste	replace the .mez

Pasting is the path that survives an IT policy, because nothing is installed at all — the reader is just part of your query. It is also the only path in Excel, which has no custom-connector loader. The .mez is the convenience path when you read these formats often in Power BI Desktop on a machine you control.

Option 1 — paste the source

Each connector is one self-contained .pq file. There are no cross-file references, deliberately: the paste is the product.

1. Open the .pq file for the connector you want from the repository and copy all of it.
2. In Power Query, New Source → Blank Query, then Advanced Editor.
3. Replace the contents with the pasted M and click Done.
4. Rename the query to exactly the function name — Sqlite3.Database, Dbf.Table, and so on. Power Query treats a query whose expression is a function as an invocable function, and the other queries find it by name.

Then call it from any other query:

let
    Source = Sqlite3.Database(File.Contents("C:\data\chinook.db")),
    Tracks = Source{[Name = "tracks"]}[Data]
in
    Tracks

Two connectors need a second query

Gpkg.Database and Mbtiles.Document read files that are SQLite databases, so they call Sqlite3.Database rather than embedding a second copy of the b-tree parser. Paste both, and name the second query exactly Sqlite3.Database. Every other connector is standalone.

Option 2 — load PQDriverless.mez

The .mez is a section document holding every connector, so all the functions are available everywhere in Power BI Desktop without pasting anything.

1. Download PQDriverless.mez (see below).
2. Copy it to [Documents]\Power BI Desktop\Custom Connectors — create the folder if it isn't there. Power BI Desktop also reads [Documents]\Microsoft Power BI Desktop\Custom Connectors, the path Microsoft's own documentation gives; either one works. If your organisation redirects Documents to OneDrive, [Documents] means the redirected folder, not %USERPROFILE%\Documents.
3. In Power BI Desktop, File → Options and settings → Options → Security → Data Extensions, and select (Not Recommended) Allow any extension to load without validation or warning.
4. Restart Power BI Desktop.

The functions are then in the global namespace and callable from a blank query, exactly as if you had pasted them:

Sqlite3.Database(File.Contents("C:\data\chinook.db"))

No Get Data entry

PQDriverless publishes M functions, not data sources. It deliberately adds nothing to the Get Data dialog and asks for no credentials of its own — the connectors take a binary that you have already fetched, so authentication is whatever File.Contents, Web.Contents or SharePoint.Files already does. Call the functions from a blank query or the formula bar.

Excel cannot load the .mez

Custom connectors are a Power BI Desktop feature. Excel's Power Query has no Custom Connectors folder and no Data Extensions setting — its Query Options → Security page offers only native database queries, certificate revocation, web preview and ADFS. Microsoft's connector extensibility documentation covers Power BI Desktop exclusively and does not mention Excel.

In Excel, paste the source instead — Option 1 above. Nothing is lost by it: each reader is one self-contained query and the M runs identically in either host.

That security setting is a real trade-off

The step 3 setting lets any unsigned extension in that folder load, not just this one. It is the standard requirement for every uncertified custom connector, and it is exactly the setting a locked-down organisation will refuse to change — which is why pasting the source exists as an option, and why a signed build is the answer for an enterprise deployment (see below).

Downloading

Grab PQDriverless.mez from the latest release. Release assets are versioned, permanently linked and downloadable without a GitHub account.

Every pull request also builds the package in CI, and the resulting PQDriverless-mez artifact is attached to the workflow run. That is the way to get an unreleased fix, with two caveats: downloading an Actions artifact requires being signed in to GitHub, and GitHub re-zips artifacts on download, so you get PQDriverless-mez.zip and must extract the .mez from it before copying it to the connectors folder.

Signed builds

The .mez above is unsigned, which is why it needs the permissive Data Extensions setting. A code-signed .pqx removes that: the publisher's certificate thumbprint is registered once by IT, after which the connector loads under the recommended security setting, with no blanket "allow any extension" policy in place.

That is the deployment story for an organisation that can't or won't loosen the setting across its estate, and it comes with versioned builds and a support contact rather than a GitHub issue queue.

If that is your situation, or you need a format that isn't in the list yet, open an issue or say hello through the repository — sponsored format development and signed builds are both available, and knowing which formats people actually need is what drives the build order.