PkgGoDev Build Status Go Report Card codebeat badge codecov


Golang commandline wrapper for wkhtmltopdf

See for wkhtmltopdf docs.

What and why

We needed a way to generate PDF documents from Go. These vary from invoices with highly customizable lay-outs to reports with tables, graphs and images. In our opinion the best way to do this was by using HTML/CSS templates as source for our PDFs. Using CSS print media types and millimeters instead of pixel units we can generate very acurate PDF documents using wkhtmltopdf.

go-wkhtmltopdf is a pure Golang wrapper around the wkhtmltopdf command line utility.

It has all options typed out as struct members which makes it very easy to use if you use an IDE with code completion and it has type safety for all options. For example you can set general options like


The same goes for adding pages, settings page options, TOC options per page etc.

It takes care of setting the correct order of options as these can become very long with muliple pages where you have page and TOC options for each page.

Secondly it makes usage in server-type applications easier, every instance (PDF process) has its own output buffer which contains the PDF output and you can feed one input document from an io.Reader (using stdin in wkhtmltopdf). You can combine any number of external HTML documents (HTTP(S) links) with at most one HTML document from stdin and set options for each input document.

Note: You can also ignore the internal buffer and let wkhtmltopdf write directly to disk if required for large files, or use the SetOutput method to pass any io.Writer.

For us this is one of the easiest ways to generate PDF documents from Go(lang) and performance is very acceptable.


go get or use a Go dependency manager of your liking.

go get -u

go-wkhtmltopdf finds the path to wkhtmltopdf by

  • first looking in the current dir
  • looking in the PATH and PATHEXT environment dirs
  • using the WKHTMLTOPDF_PATH environment dir

If you need to set your own wkhtmltopdf path or want to change it during execution, you can call SetPath().


See testfile wkhtmltopdf_test.go for more complex options, a common use case test is in simplesample_test.go

package wkhtmltopdf

import (

func ExampleNewPDFGenerator() {

  // Create new PDF generator
  pdfg, err := NewPDFGenerator()
  if err != nil {

  // Set global options

  // Create a new input page from an URL
  page := NewPage("")

  // Set options for this page

  // Add to document

  // Create PDF document in internal buffer
  err = pdfg.Create()
  if err != nil {

  // Write buffer contents to file on disk
  err = pdfg.WriteFile("./simplesample.pdf")
  if err != nil {

  // Output: Done

As mentioned before, you can provide one document from stdin, this is done by using a PageReader object as input to AddPage. This is best constructed with NewPageReader and will accept any io.Reader so this can be used with files from disk (os.File) or memory (bytes.Buffer) etc.
A simple example snippet:

html := "<html>Hi</html>"

Saving to and loading from JSON

The package now has the possibility to save the PDF Generator object as JSON and to create a new PDF Generator from a JSON file. All options and pages are saved in JSON, pages added using NewPageReader are read to memory before saving and then saved as Base64 encoded strings in the JSON file.

This is useful to prepare a PDF file and generate the actual PDF elsewhere, for example on AWS Lambda. To create PDF Generator on the client, where wkhtmltopdf might not be present, function NewPDFPreparer can be used.

Use NewPDFPreparer to create a PDF Generator object on the client and NewPDFGeneratorFromJSON to reconstruct it on the server.

// Client code
pdfg := NewPDFPreparer()
htmlfile, err := ioutil.ReadFile("./testfiles/htmlsimple.html")
if err != nil {
// The contents of htmlsimple.html are saved as base64 string in the JSON file
jb, err := pdfg.ToJSON()
if err != nil {
// Server code
pdfgFromJSON, err := NewPDFGeneratorFromJSON(bytes.NewReader(jb))
if err != nil {
err = pdfgFromJSON.Create()
if err != nil {

For an example of running this in AWS Lambda see


The speed if pretty much determined by wkhtmltopdf itself, or if you use external source URLs, the time it takes to get and render the source HTML.

The go wrapper time is negligible with around 0.04ms for parsing an above average number of commandline options.

Benchmarks are included.

Expand ▾ Collapse ▴



    Package wkhtmltopdf contains wrappers around the wkhtmltopdf commandline tool




    View Source
    const (
    	OrientationLandscape = "Landscape" // Landscape mode
    	OrientationPortrait  = "Portrait"  // Portrait mode

      Constants for orientation modes

      View Source
      const (
      	PageSizeA0        = "A0"        //	841 x 1189 mm
      	PageSizeA1        = "A1"        //	594 x 841 mm
      	PageSizeA2        = "A2"        //	420 x 594 mm
      	PageSizeA3        = "A3"        //	297 x 420 mm
      	PageSizeA4        = "A4"        //	210 x 297 mm, 8.26
      	PageSizeA5        = "A5"        //	148 x 210 mm
      	PageSizeA6        = "A6"        //	105 x 148 mm
      	PageSizeA7        = "A7"        //	74 x 105 mm
      	PageSizeA8        = "A8"        //	52 x 74 mm
      	PageSizeA9        = "A9"        //	37 x 52 mm
      	PageSizeB0        = "B0"        //	1000 x 1414 mm
      	PageSizeB1        = "B1"        //	707 x 1000 mm
      	PageSizeB10       = "B10"       //	31 x 44 mm
      	PageSizeB2        = "B2"        //	500 x 707 mm
      	PageSizeB3        = "B3"        //	353 x 500 mm
      	PageSizeB4        = "B4"        //	250 x 353 mm
      	PageSizeB5        = "B5"        //	176 x 250 mm, 6.93
      	PageSizeB6        = "B6"        //	125 x 176 mm
      	PageSizeB7        = "B7"        //	88 x 125 mm
      	PageSizeB8        = "B8"        //	62 x 88 mm
      	PageSizeB9        = "B9"        //	33 x 62 mm
      	PageSizeC5E       = "C5E"       //	163 x 229 mm
      	PageSizeComm10E   = "Comm10E"   //	105 x 241 mm, U.S. Common 10 Envelope
      	PageSizeCustom    = "Custom"    //	Unknown, or a user defined size.
      	PageSizeDLE       = "DLE"       //	110 x 220 mm
      	PageSizeExecutive = "Executive" //	7.5 x 10 inches, 190.5 x 254 mm
      	PageSizeFolio     = "Folio"     //	210 x 330 mm
      	PageSizeLedger    = "Ledger"    //	431.8 x 279.4 mm
      	PageSizeLegal     = "Legal"     //	8.5 x 14 inches, 215.9 x 355.6 mm
      	PageSizeLetter    = "Letter"    //	8.5 x 11 inches, 215.9 x 279.4 mm
      	PageSizeTabloid   = "Tabloid"   //	279.4 x 431.8 mm

        Constants for page sizes


        This section is empty.


        func GetPath

        func GetPath() string

          GetPath gets the path to wkhtmltopdf

          func SetPath

          func SetPath(path string)

            SetPath sets the path to wkhtmltopdf


            type PDFGenerator

            type PDFGenerator struct {
            	Cover      cover
            	TOC        toc
            	OutputFile string //filename to write to, default empty (writes to internal buffer)
            	// contains filtered or unexported fields

              PDFGenerator is the main wkhtmltopdf struct, always use NewPDFGenerator to obtain a new PDFGenerator struct

              func NewPDFGenerator

              func NewPDFGenerator() (*PDFGenerator, error)

                NewPDFGenerator returns a new PDFGenerator struct with all options created and checks if wkhtmltopdf can be found on the system


                func NewPDFGeneratorFromJSON

                func NewPDFGeneratorFromJSON(jsonReader io.Reader) (*PDFGenerator, error)

                  NewPDFGeneratorFromJSON creates a new PDFGenerator and restores all the settings and pages from a JSON byte slice which should be created using PDFGenerator.ToJSON().


                  func NewPDFPreparer

                  func NewPDFPreparer() *PDFGenerator

                    NewPDFPreparer returns a PDFGenerator object without looking for the wkhtmltopdf executable file. This is useful to prepare a PDF file that is generated elsewhere and you just want to save the config as JSON. Note that Create() can not be called on this object unless you call SetPath yourself.

                    func (*PDFGenerator) AddPage

                    func (pdfg *PDFGenerator) AddPage(p PageProvider)

                      AddPage adds a new input page to the document. A page is an input HTML page, it can span multiple pages in the output document. It is a Page when read from file or URL or a PageReader when read from memory.

                      func (*PDFGenerator) ArgString

                      func (pdfg *PDFGenerator) ArgString() string

                        ArgString returns Args as a single string

                        func (*PDFGenerator) Args

                        func (pdfg *PDFGenerator) Args() []string

                          Args returns the commandline arguments as a string slice

                          func (*PDFGenerator) Buffer

                          func (pdfg *PDFGenerator) Buffer() *bytes.Buffer

                            Buffer returns the embedded output buffer used if OutputFile is empty

                            func (*PDFGenerator) Bytes

                            func (pdfg *PDFGenerator) Bytes() []byte

                              Bytes returns the output byte slice from the output buffer used if OutputFile is empty

                              func (*PDFGenerator) Create

                              func (pdfg *PDFGenerator) Create() error

                                Create creates the PDF document and stores it in the internal buffer if no error is returned

                                func (*PDFGenerator) ResetPages

                                func (pdfg *PDFGenerator) ResetPages()

                                  ResetPages drops all pages previously added by AddPage or SetPages. This allows reuse of current instance of PDFGenerator with all of it's configuration preserved.

                                  func (*PDFGenerator) SetOutput

                                  func (pdfg *PDFGenerator) SetOutput(w io.Writer)

                                    SetOutput sets the output to write the PDF to, when this method is called, the internal buffer will not be used, so the Bytes(), Buffer() and WriteFile() methods will not work.

                                    func (*PDFGenerator) SetPages

                                    func (pdfg *PDFGenerator) SetPages(p []PageProvider)

                                      SetPages resets all pages

                                      func (*PDFGenerator) SetStderr

                                      func (pdfg *PDFGenerator) SetStderr(w io.Writer)

                                        SetStderr sets the output writer for Stderr when running the wkhtmltopdf command. You only need to call this when you want to print the output of wkhtmltopdf (like the progress messages in verbose mode). If not called, or if w is nil, the output of Stderr is kept in an internal buffer and returned as error message if there was an error when calling wkhtmltopdf.

                                        func (*PDFGenerator) ToJSON

                                        func (pdfg *PDFGenerator) ToJSON() ([]byte, error)

                                          ToJSON creates JSON of the complete representation of the PDFGenerator. It also saves all pages. For a PageReader page, the content is stored as a Base64 string in the JSON.

                                          func (*PDFGenerator) WriteFile

                                          func (pdfg *PDFGenerator) WriteFile(filename string) error

                                            WriteFile writes the contents of the output buffer to a file

                                            type Page

                                            type Page struct {
                                            	Input string

                                              Page is the input struct for each page

                                              func NewPage

                                              func NewPage(input string) *Page

                                                NewPage creates a new input page from a local or web resource (filepath or URL)

                                                func (*Page) Args

                                                func (p *Page) Args() []string

                                                  Args returns the argument slice and is part of the page interface

                                                  func (*Page) InputFile

                                                  func (p *Page) InputFile() string

                                                    InputFile returns the input string and is part of the page interface

                                                    func (*Page) Reader

                                                    func (p *Page) Reader() io.Reader

                                                      Reader returns the io.Reader and is part of the page interface

                                                      type PageOptions

                                                      type PageOptions struct {
                                                      	// contains filtered or unexported fields

                                                        PageOptions are options for each input page

                                                        func NewPageOptions

                                                        func NewPageOptions() PageOptions

                                                          NewPageOptions returns a new PageOptions struct with all options

                                                          func (*PageOptions) Args

                                                          func (po *PageOptions) Args() []string

                                                            Args returns the argument slice

                                                            type PageProvider

                                                            type PageProvider interface {
                                                            	Args() []string
                                                            	InputFile() string
                                                            	Reader() io.Reader

                                                              PageProvider is the interface which provides a single input page. Implemented by Page and PageReader.

                                                              type PageReader

                                                              type PageReader struct {
                                                              	Input io.Reader

                                                                PageReader is one input page (a HTML document) that is read from an io.Reader You can add only one Page from a reader

                                                                func NewPageReader

                                                                func NewPageReader(input io.Reader) *PageReader

                                                                  NewPageReader creates a new PageReader from an io.Reader

                                                                  func (*PageReader) Args

                                                                  func (pr *PageReader) Args() []string

                                                                    Args returns the argument slice and is part of the page interface

                                                                    func (*PageReader) InputFile

                                                                    func (pr *PageReader) InputFile() string

                                                                      InputFile returns the input string and is part of the page interface

                                                                      func (*PageReader) Reader

                                                                      func (pr *PageReader) Reader() io.Reader

                                                                        Reader returns the io.Reader and is part of the page interface