diff --git a/lib/cmds/files_cmds/download.js b/lib/cmds/files_cmds/download.js
index e95e10e..a0a6589 100644
--- a/lib/cmds/files_cmds/download.js
+++ b/lib/cmds/files_cmds/download.js
@@ -1,22 +1,52 @@
 'use strict';
 
+const querystring = require('querystring');
 const path = require('path');
-const { get, download } = require('../../api');
+const { get, download, list } = require('../../api');
 
-exports.command = 'download <fileId> [destDir]';
-exports.desc = 'Download file content by <fileId>';
+exports.command = 'download <source> [destDir]';
+exports.desc = 'Download file content by <source>';
 exports.builder = yargs => {
-  yargs.positional('fileId', {
-    describe: 'The ID of the file to download.',
+  yargs.positional('source', {
+    describe: 'Either the ID of the file to download or a project ID and prefix of the form <project ID>/<prefix> to download a set of files from a project.',
     type: 'string'
   }).positional('destDir', {
     describe: 'The destination directory for the download file.  Defaults to current working directory',
     type: 'string'
+  }).option('recursive', {
+    describe:
+      'If <source> is a project Id with an optional prefix, download all files under that prefix to the destination.',
+    alias: 'r',
+    type: 'boolean',
+    default: false
+  }).option('limit', {
+    describe: 'The maximum number of files to download when used with the --recursive option',
+    alias: 'l',
+    type: 'number',
+    default: 1000
   });
 };
 
 exports.handler = async argv => {
-  const response = await get(argv, `/v1/files/${argv.fileId}`);
+  if (argv.recursive) {
+    const [datasetId, prefix] = argv.source.split('/');
 
-  await download(argv, `/v1/files/${argv.fileId}?include=downloadUrl`, path.join(argv.destDir || process.cwd(), response.data.name));
+    const opts = {
+      datasetId,
+      pageSize: argv.limit
+    };
+
+    if (prefix) {
+      opts.name = prefix;
+    }
+
+    const response = await list(argv, `/v1/files?${querystring.stringify(opts)}`);
+    for (const item of response.data.items) {
+      await download(argv, `/v1/files/${item.id}?include=downloadUrl`, path.join(argv.destDir || process.cwd(), item.name));
+    }
+  } else {
+    const response = await get(argv, `/v1/files/${argv.source}`);
+
+    await download(argv, `/v1/files/${argv.source}?include=downloadUrl`, path.join(argv.destDir || process.cwd(), response.data.name));
+  }
 };
diff --git a/test/unit/commands/file.test.js b/test/unit/commands/file.test.js
index 19fdb3e..66093f7 100644
--- a/test/unit/commands/file.test.js
+++ b/test/unit/commands/file.test.js
@@ -207,6 +207,35 @@ test.serial.cb('The "files-download" command should download a file', t => {
     .parse('download fileid /dir');
 });
 
+test.serial.cb('The "files-download" command should download a set of files from a project', t => {
+  listStub.onFirstCall().returns({
+    data: {
+      items: [
+        {
+          id: '1',
+          name: 'foo.txt'
+        }
+      ]
+    }
+  });
+
+  callback = () => {
+    t.is(listStub.callCount, 1);
+    t.is(listStub.getCall(0).args[1], '/v1/files?datasetId=projectId&pageSize=1000&name=prefix');
+
+    t.is(downloadSpy.callCount, 1);
+    t.is(downloadSpy.getCall(0).args[1], '/v1/files/1?include=downloadUrl');
+    t.is(downloadSpy.getCall(0).args[2], '/dir/foo.txt');
+    t.end();
+  };
+
+  t.context.deleteFileStub = t.context.sandbox.stub(fs, 'unlinkSync').callsFake(callback);
+  t.context.copyFileStub = t.context.sandbox.stub(fs, 'copyFileSync').callsFake(callback);
+
+  yargs.command(download)
+    .parse('download projectId/prefix /dir -r');
+});
+
 test.serial.cb('The "files-upload" command should upload a file', t => {
   const res = { data: { uploadUrl: 'https://host/upload' } };
   postStub.onFirstCall().returns(res);