diff --git a/.gitignore b/.gitignore
index 6550409..8f73e30 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,2 +1,2 @@
-dl
+./dl
 *.part*
diff --git a/cmd/dl/dl_part.go b/cmd/dl/dl_part.go
new file mode 100644
index 0000000..c1eb430
--- /dev/null
+++ b/cmd/dl/dl_part.go
@@ -0,0 +1,52 @@
+package main
+
+import (
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"path"
+	"sync"
+
+	"github.com/schollz/progressbar/v3"
+)
+
+type downloadPart struct {
+	index     int
+	uri       string
+	dir       string
+	filename  string
+	startByte uint64
+	endByte   uint64
+}
+
+func (p *downloadPart) downloadPartFilename() string {
+	return path.Join(p.dir, fmt.Sprintf("download.part%d", p.index))
+}
+
+func (p *downloadPart) fetchPart(wg *sync.WaitGroup, bar *progressbar.ProgressBar) {
+	defer wg.Done()
+
+	byteRange := fmt.Sprintf("bytes=%d-%d", p.startByte, p.endByte)
+	req, _ := http.NewRequest("GET", p.uri, nil)
+	req.Header.Set("Range", byteRange)
+	req.Header.Set("User-Agent", "dl/1.0")
+
+	client := &http.Client{}
+	resp, err := client.Do(req)
+	if err != nil {
+		return
+	}
+	defer resp.Body.Close()
+
+	// Create the file
+	filename := p.downloadPartFilename()
+	out, err := os.Create(filename)
+	if err != nil {
+		return
+	}
+	defer out.Close()
+
+	// Write the body to file
+	_, _ = io.Copy(io.MultiWriter(out, bar), resp.Body)
+}
diff --git a/cmd/dl/main.go b/cmd/dl/main.go
new file mode 100644
index 0000000..39b3fa0
--- /dev/null
+++ b/cmd/dl/main.go
@@ -0,0 +1,203 @@
+package main
+
+import (
+	"flag"
+	"fmt"
+	"io"
+	"mime"
+	"net/http"
+	"os"
+	"os/signal"
+	"strconv"
+	"strings"
+	"sync"
+	"syscall"
+
+	"github.com/schollz/progressbar/v3"
+)
+
+type download struct {
+	uri        string
+	filesize   uint64
+	filename   string
+	workingDir string
+	boost      int
+	parts      []downloadPart
+}
+
+func main() {
+	filenamePtr := flag.String("filename", "", "custom filename")
+	boostPtr := flag.Int("boost", 8, "number of concurrent downloads")
+	workingDirPtr := flag.String("workdir", "", "working directory for downloads")
+
+	flag.Parse()
+
+	file_uris := flag.Args()
+
+	var err error
+
+	for _, uri := range file_uris {
+		var dl download
+		dl.uri = uri
+		dl.boost = *boostPtr
+
+		err = dl.FetchMetadata()
+		if err != nil {
+			panic(err)
+		}
+
+		// Use filename from args if specified
+		if *filenamePtr != "" {
+			dl.filename = *filenamePtr
+		}
+
+		if *workingDirPtr != "" {
+			dl.workingDir = *workingDirPtr
+		} else {
+			dl.workingDir, err = os.Getwd()
+			if err != nil {
+				panic(err)
+			}
+		}
+
+		sigc := make(chan os.Signal, 1)
+		signal.Notify(sigc,
+			syscall.SIGHUP,
+			syscall.SIGINT,
+			syscall.SIGTERM,
+			syscall.SIGQUIT)
+		go func() {
+			sig := <-sigc
+			fmt.Printf("\n%s; cleaning up...\n", sig)
+			dl.cleanupParts()
+			os.Exit(0)
+		}()
+
+		fmt.Println(dl.filename)
+
+		dl.Fetch()
+		dl.ConcatFiles()
+	}
+}
+
+func (dl *download) FetchMetadata() error {
+	resp, err := http.Head(dl.uri)
+	if err != nil {
+		return err
+	}
+	defer resp.Body.Close()
+
+	contentLength := resp.Header.Get("Content-Length")
+	dl.filesize, err = strconv.ParseUint(contentLength, 0, 64)
+	if err != nil {
+		return err
+	}
+
+	contentDisposition := resp.Header.Get("Content-Disposition")
+	_, params, err := mime.ParseMediaType(contentDisposition)
+	if err != nil {
+		dl.filename = dl.filenameFromURI()
+		return err
+	} else {
+		dl.filename = params["filename"]
+	}
+
+	// No filename specified in the header; use the pathname
+	if dl.filename == "" {
+		dl.filename = dl.filenameFromURI()
+	}
+
+	return nil
+}
+
+func (dl *download) Fetch() error {
+	var wg sync.WaitGroup
+
+	bar := progressbar.DefaultBytes(
+		int64(dl.filesize),
+		"Downloading",
+	)
+
+	for i := 0; i < dl.boost; i++ {
+		start, end := dl.calculatePartBoundary(i)
+		wg.Add(1)
+		dlPart := downloadPart{
+			index:     i,
+			uri:       dl.uri,
+			dir:       dl.workingDir,
+			startByte: start,
+			endByte:   end,
+		}
+		dlPart.filename = dlPart.downloadPartFilename()
+		dl.parts = append(dl.parts, dlPart)
+		go dlPart.fetchPart(&wg, bar)
+	}
+
+	wg.Wait()
+	return nil
+}
+
+func (dl *download) calculatePartBoundary(part int) (startByte uint64, endByte uint64) {
+	chunkSize := dl.filesize / uint64(dl.boost)
+	var previousEndByte uint64
+
+	if part == 0 {
+		startByte = 0
+		previousEndByte = 0
+	} else {
+		previousEndByte = uint64(part)*chunkSize - 1
+		startByte = previousEndByte + 1
+	}
+
+	// For the last part, pick up all remaining bytes
+	if part == (dl.boost - 1) {
+		endByte = dl.filesize - 1
+	} else {
+		endByte = startByte + chunkSize - 1
+	}
+
+	return
+}
+
+func (dl *download) filenameFromURI() string {
+	splitURI := strings.Split(dl.uri, "/")
+	return splitURI[len(splitURI)-1]
+}
+
+func (dl *download) ConcatFiles() {
+	var readers []io.Reader
+
+	defer dl.cleanupParts()
+
+	bar := progressbar.DefaultBytes(
+		int64(dl.filesize),
+		"Combining  ",
+	)
+
+	for _, part := range dl.parts {
+		downloadPart, err := os.Open(part.downloadPartFilename())
+		if err != nil {
+			panic(err)
+		}
+		defer downloadPart.Close()
+		readers = append(readers, downloadPart)
+	}
+
+	inputFiles := io.MultiReader(readers...)
+
+	outFile, err := os.Create(dl.filename)
+	if err != nil {
+		panic(err)
+	}
+
+	_, err = io.Copy(io.MultiWriter(outFile, bar), inputFiles)
+	if err != nil {
+		panic(err)
+	}
+}
+
+func (dl *download) cleanupParts() {
+	for _, part := range dl.parts {
+		os.Remove(part.downloadPartFilename())
+	}
+}