diff --git a/README.md b/README.md
index 1646965..9097605 100644
--- a/README.md
+++ b/README.md
@@ -47,6 +47,13 @@ pdfops stamp -text "seen 25 August" -at top-right -bold notes.pdf seen.pdf
 pdfops sanitize downloaded.pdf safe.pdf
 pdfops flatten filled-form.pdf final.pdf
 pdfops strip -annotations -bookmarks private.pdf clean.pdf
+pdfops compress fat.pdf lean.pdf
+pdfops encrypt -user letmein -allow print,copy plain.pdf locked.pdf
+pdfops -password letmein decrypt locked.pdf plain.pdf
+pdfops -password letmein permissions locked.pdf
+pdfops text paper.pdf
+pdfops text -layout -pages 1 paper.pdf
+pdfops images paper.pdf pictures/
 pdfops info file.pdf
 ```
 
@@ -54,7 +61,13 @@ A page range is written `1-3,7,10-` and may say `all`, `even`, `odd` or
 `last`. It keeps its own order and its own repeats, so `select -pages 3-1`
 reverses three pages and `select -pages 1,1` gives you two copies.
 
-`-password` opens an encrypted file; what is written out is not encrypted.
+`-password` opens an encrypted file.
+
+`text` reads the page back as words, and `-layout` says where every piece of
+it sits — a page, a place, a size, and what it says. A piece the document
+gave no way to read comes back marked rather than guessed at, so you can tell
+a page that says nothing from one that could not be read. `images` writes out
+the pictures a page places, a JPEG as a JPEG.
 
 Links and bookmarks are carried over and pointed at the pages they became
 here, so extracting three pages of a book leaves the links between those
diff --git a/cmd/pdfops/run.go b/cmd/pdfops/run.go
index 40bda93..5f3dad8 100644
--- a/cmd/pdfops/run.go
+++ b/cmd/pdfops/run.go
@@ -10,6 +10,7 @@ import (
 	"strconv"
 	"strings"
 
+	"github.com/go-pdfkit/extract"
 	"github.com/go-pdfkit/ops"
 	"github.com/go-pdfkit/reader"
 )
@@ -55,6 +56,8 @@ var commands = []command{
 	{"encrypt", "-user <password> [-owner <password>] [-allow <what>] [-aes128] <in.pdf> <out.pdf>", "protect the file with a password", runEncrypt},
 	{"decrypt", "<in.pdf> <out.pdf>", "write the file without its protection", runDecrypt},
 	{"permissions", "<in.pdf>", "say how the file is protected and what it allows", runPermissions},
+	{"text", "[-pages <range>] [-layout] <in.pdf>", "read the text off the pages", runText},
+	{"images", "[-pages <range>] <in.pdf> <out-directory>", "write out the pictures the pages place", runImages},
 }
 
 // run is the whole program, so that the tests can drive it.
@@ -778,3 +781,121 @@ func openedAs(owner bool) string {
 	}
 	return "the user"
 }
+
+// runText reads the text off a document and prints it.
+func runText(c *context, args []string) error {
+	fs := flags("text")
+	spec := fs.String("pages", "all", "which pages to read")
+	layout := fs.Bool("layout", false, "print where each piece of text sits as well as what it says")
+	if err := fs.Parse(args); err != nil {
+		return err
+	}
+	if err := wantArgs(fs, 1, "<in.pdf>"); err != nil {
+		return err
+	}
+	src, err := c.read(fs.Arg(0))
+	if err != nil {
+		return err
+	}
+	pages, err := ops.ParseRange(*spec, src.PageCount())
+	if err != nil {
+		return err
+	}
+	for _, page := range pages {
+		// Every page here came from the range, so it is one the document
+		// has; neither of these can fail.
+		if *layout {
+			runs, _ := extract.Runs(src, page)
+			for _, r := range runs {
+				fmt.Fprintf(c.out, "%d\t%.2f\t%.2f\t%.2f\t%s%s\n",
+					page, r.X, r.Y, r.Size, marks(r), r.Text)
+			}
+			continue
+		}
+		text, _ := extract.Text(src, page)
+		fmt.Fprintln(c.out, text)
+	}
+	return nil
+}
+
+// marks says what is unusual about a run, in front of what it says.
+func marks(r extract.Run) string {
+	switch {
+	case r.Invisible && r.Unreadable:
+		return "[invisible, part unreadable] "
+	case r.Invisible:
+		return "[invisible] "
+	case r.Unreadable:
+		return "[part unreadable] "
+	}
+	return ""
+}
+
+// runImages writes out the pictures a document places.
+func runImages(c *context, args []string) error {
+	fs := flags("images")
+	spec := fs.String("pages", "all", "which pages to take pictures from")
+	if err := fs.Parse(args); err != nil {
+		return err
+	}
+	if err := wantArgs(fs, 2, "<in.pdf> <out-directory>"); err != nil {
+		return err
+	}
+	src, err := c.read(fs.Arg(0))
+	if err != nil {
+		return err
+	}
+	pages, err := ops.ParseRange(*spec, src.PageCount())
+	if err != nil {
+		return err
+	}
+	dir := fs.Arg(1)
+	if err := os.MkdirAll(dir, 0o755); err != nil {
+		return err
+	}
+	n := 0
+	for _, page := range pages {
+		// The range only names pages the document has.
+		images, _ := extract.Images(src, page)
+		for i, im := range images {
+			name := fmt.Sprintf("page%03d-%02d%s", page, i+1, imageSuffix(im))
+			path := filepath.Join(dir, name)
+			if err := os.WriteFile(path, im.Data, 0o644); err != nil {
+				return err
+			}
+			fmt.Fprintf(c.out, "%s\t%dx%d\tat %.1f,%.1f\tdrawn %.1fx%.1f\n",
+				name, im.Width, im.Height, im.X, im.Y, im.DrawnWidth, im.DrawnHeight)
+			n++
+		}
+	}
+	if n == 0 {
+		fmt.Fprintln(c.out, "no pictures")
+	}
+	return nil
+}
+
+// imageSuffix names a picture by what it holds. A JPEG is written out as one;
+// anything this has unfiltered into plain samples is written as it stands,
+// since turning samples into pixels means reading a colour space and that is
+// the renderer's work.
+func imageSuffix(im extract.Image) string {
+	switch im.Filter {
+	case "DCTDecode":
+		return ".jpg"
+	case "JPXDecode":
+		return ".jp2"
+	case "JBIG2Decode":
+		return ".jbig2"
+	}
+	return ".samples"
+}
+
+// read opens a document rather than a document being assembled, which is what
+// reading a page back needs.
+func (c *context) read(path string) (*reader.Document, error) {
+	b, err := os.ReadFile(path)
+	if err != nil {
+		return nil, err
+	}
+	return reader.OpenWithPassword(b, c.password)
+}
diff --git a/cmd/pdfops/run_test.go b/cmd/pdfops/run_test.go
index 5789082..f9bf5db 100644
--- a/cmd/pdfops/run_test.go
+++ b/cmd/pdfops/run_test.go
@@ -8,6 +8,7 @@ import (
 	"strings"
 	"testing"
 
+	"github.com/go-pdfkit/extract"
 	"github.com/go-pdfkit/ops"
 	"github.com/go-pdfkit/reader"
 )
@@ -636,3 +637,212 @@ func TestTheVerbsThatProtectRefuseWhatTheyShould(t *testing.T) {
 		}
 	}
 }
+
+// pageWithText writes a document whose page carries real text and a picture,
+// so that reading it back can be checked.
+func pageWithText(t *testing.T) string {
+	t.Helper()
+	w := reader.NewWriter("1.7")
+	pagesRef := w.Reserve()
+	widths := make(reader.Array, 0, 224)
+	for i := 32; i < 256; i++ {
+		widths = append(widths, reader.Integer(500))
+	}
+	font := w.Add(reader.Dict{
+		"Type": reader.Name("Font"), "Subtype": reader.Name("Type1"),
+		"BaseFont": reader.Name("Helvetica"), "FirstChar": reader.Integer(32),
+		"LastChar": reader.Integer(255), "Widths": widths,
+		"Encoding": reader.Name("WinAnsiEncoding"),
+	})
+	image := w.Add(&reader.Stream{Dict: reader.Dict{
+		"Type": reader.Name("XObject"), "Subtype": reader.Name("Image"),
+		"Width": reader.Integer(2), "Height": reader.Integer(2),
+		"ColorSpace": reader.Name("DeviceGray"), "BitsPerComponent": reader.Integer(8),
+	}, Raw: []byte{0, 64, 128, 255}})
+	jpeg := w.Add(&reader.Stream{Dict: reader.Dict{
+		"Subtype": reader.Name("Image"), "Width": reader.Integer(1),
+		"Height": reader.Integer(1), "Filter": reader.Name("DCTDecode"),
+	}, Raw: []byte("pretend")})
+	kids := make(reader.Array, 0, 2)
+	for i := 1; i <= 2; i++ {
+		content := fmt.Sprintf("BT /F1 12 Tf 20 100 Td (page %d says hello) Tj ET "+
+			"q 20 0 0 20 10 10 cm /Im Do Q q 5 0 0 5 50 50 cm /Jp Do Q", i)
+		kids = append(kids, w.Add(reader.Dict{
+			"Type": reader.Name("Page"), "Parent": pagesRef,
+			"MediaBox": reader.Array{reader.Integer(0), reader.Integer(0),
+				reader.Integer(200), reader.Integer(200)},
+			"Contents": w.Add(&reader.Stream{Dict: reader.Dict{}, Raw: []byte(content)}),
+			"Resources": reader.Dict{
+				"Font":    reader.Dict{"F1": font},
+				"XObject": reader.Dict{"Im": image, "Jp": jpeg},
+			},
+		}))
+	}
+	w.Put(pagesRef, reader.Dict{"Type": reader.Name("Pages"),
+		"Kids": kids, "Count": reader.Integer(len(kids))})
+	root := w.Add(reader.Dict{"Type": reader.Name("Catalog"), "Pages": pagesRef})
+	out, err := w.Finish(reader.Dict{"Root": root})
+	if err != nil {
+		t.Fatal(err)
+	}
+	path := filepath.Join(t.TempDir(), "text.pdf")
+	if err := os.WriteFile(path, out, 0o644); err != nil {
+		t.Fatal(err)
+	}
+	return path
+}
+
+func TestTextVerb(t *testing.T) {
+	in := pageWithText(t)
+	code, out, errOut := exec("text", in)
+	if code != 0 {
+		t.Fatalf("text said %d: %s", code, errOut)
+	}
+	for _, want := range []string{"page 1 says hello", "page 2 says hello"} {
+		if !strings.Contains(out, want) {
+			t.Errorf("the text does not hold %q:\n%s", want, out)
+		}
+	}
+	if _, out, _ = exec("text", "-pages", "2", in); strings.Contains(out, "page 1") {
+		t.Errorf("a range of one page gave:\n%s", out)
+	}
+	// The layout form says where every piece sits.
+	code, out, errOut = exec("text", "-layout", "-pages", "1", in)
+	if code != 0 {
+		t.Fatalf("text -layout said %d: %s", code, errOut)
+	}
+	if !strings.Contains(out, "\t20.00\t100.00\t12.00\t") {
+		t.Errorf("the layout does not say where the text is:\n%s", out)
+	}
+	for _, args := range [][]string{
+		{"text"},
+		{"text", "nowhere.pdf"},
+		{"text", "-nope", in},
+		{"text", "-pages", "nonsense", in},
+	} {
+		if code, _, _ := exec(args...); code == 0 {
+			t.Errorf("%v was accepted", args)
+		}
+	}
+}
+
+func TestTextVerbSaysWhatItCouldNotRead(t *testing.T) {
+	// A run nothing in the document could name is marked rather than left
+	// out, and so is one drawn with no ink.
+	w := reader.NewWriter("1.7")
+	pagesRef := w.Reserve()
+	font := w.Add(reader.Dict{
+		"Type": reader.Name("Font"), "Subtype": reader.Name("Type1"),
+		"FontDescriptor": w.Add(reader.Dict{"Flags": reader.Integer(4)}),
+	})
+	page := w.Add(reader.Dict{
+		"Type": reader.Name("Page"), "Parent": pagesRef,
+		"MediaBox": reader.Array{reader.Integer(0), reader.Integer(0),
+			reader.Integer(200), reader.Integer(200)},
+		"Contents": w.Add(&reader.Stream{Dict: reader.Dict{},
+			Raw: []byte("BT 3 Tr /F1 12 Tf 20 100 Td (\x01\x02) Tj ET")}),
+		"Resources": reader.Dict{"Font": reader.Dict{"F1": font}},
+	})
+	w.Put(pagesRef, reader.Dict{"Type": reader.Name("Pages"),
+		"Kids": reader.Array{page}, "Count": reader.Integer(1)})
+	root := w.Add(reader.Dict{"Type": reader.Name("Catalog"), "Pages": pagesRef})
+	out, err := w.Finish(reader.Dict{"Root": root})
+	if err != nil {
+		t.Fatal(err)
+	}
+	path := filepath.Join(t.TempDir(), "unreadable.pdf")
+	if err := os.WriteFile(path, out, 0o644); err != nil {
+		t.Fatal(err)
+	}
+	_, printed, _ := exec("text", "-layout", path)
+	if !strings.Contains(printed, "[invisible, part unreadable]") {
+		t.Errorf("the run was not marked:\n%s", printed)
+	}
+}
+
+func TestImagesVerb(t *testing.T) {
+	in := pageWithText(t)
+	dir := filepath.Join(t.TempDir(), "pictures")
+	code, out, errOut := exec("images", in, dir)
+	if code != 0 {
+		t.Fatalf("images said %d: %s", code, errOut)
+	}
+	if !strings.Contains(out, "page001-01.samples") || !strings.Contains(out, "page001-02.jpg") {
+		t.Errorf("the pictures are:\n%s", out)
+	}
+	files, err := os.ReadDir(dir)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(files) != 4 {
+		t.Errorf("%d files were written", len(files))
+	}
+	jpg, err := os.ReadFile(filepath.Join(dir, "page001-02.jpg"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if string(jpg) != "pretend" {
+		t.Errorf("the JPEG was changed on the way out: %q", jpg)
+	}
+	// A document with no pictures says so.
+	plain := fixture(t, 1)
+	if _, out, _ = exec("images", plain, filepath.Join(t.TempDir(), "none")); !strings.Contains(out, "no pictures") {
+		t.Errorf("a document with no pictures said:\n%s", out)
+	}
+	for _, args := range [][]string{
+		{"images", in},
+		{"images", "nowhere.pdf", dir},
+		{"images", "-nope", in, dir},
+		{"images", "-pages", "nonsense", in, dir},
+		{"images", in, "/dev/null/cannot"},
+	} {
+		if code, _, _ := exec(args...); code == 0 {
+			t.Errorf("%v was accepted", args)
+		}
+	}
+}
+
+func TestHowARunIsMarked(t *testing.T) {
+	for _, c := range []struct {
+		run  extract.Run
+		want string
+	}{
+		{extract.Run{}, ""},
+		{extract.Run{Invisible: true}, "[invisible] "},
+		{extract.Run{Unreadable: true}, "[part unreadable] "},
+		{extract.Run{Invisible: true, Unreadable: true}, "[invisible, part unreadable] "},
+	} {
+		if got := marks(c.run); got != c.want {
+			t.Errorf("%+v is marked %q, want %q", c.run, got, c.want)
+		}
+	}
+}
+
+func TestAPictureThatCannotBeWritten(t *testing.T) {
+	// Somewhere to put the pictures where one of them cannot go: a directory
+	// already standing where a file should be written.
+	in := pageWithText(t)
+	dir := filepath.Join(t.TempDir(), "pictures")
+	if err := os.MkdirAll(filepath.Join(dir, "page001-01.samples"), 0o755); err != nil {
+		t.Fatal(err)
+	}
+	if code, _, _ := exec("images", in, dir); code == 0 {
+		t.Error("writing over a directory was accepted")
+	}
+}
+
+func TestEveryWayAPictureIsNamed(t *testing.T) {
+	for _, c := range []struct {
+		filter reader.Name
+		want   string
+	}{
+		{"DCTDecode", ".jpg"},
+		{"JPXDecode", ".jp2"},
+		{"JBIG2Decode", ".jbig2"},
+		{"", ".samples"},
+	} {
+		if got := imageSuffix(extract.Image{Filter: c.filter}); got != c.want {
+			t.Errorf("%q is written as %q, want %q", c.filter, got, c.want)
+		}
+	}
+}
diff --git a/go.mod b/go.mod
index 5224f40..2719c57 100644
--- a/go.mod
+++ b/go.mod
@@ -2,4 +2,12 @@ module github.com/go-pdfkit/ops
 
 go 1.26.4
 
-require github.com/go-pdfkit/reader v0.4.0
+require (
+	github.com/go-pdfkit/extract v0.1.0
+	github.com/go-pdfkit/reader v0.4.0
+)
+
+require (
+	github.com/go-opentype/opentype v0.9.0 // indirect
+	github.com/go-pdfkit/pdffont v0.1.0 // indirect
+)
diff --git a/go.sum b/go.sum
index 3ff081a..0244ea4 100644
--- a/go.sum
+++ b/go.sum
@@ -1,2 +1,10 @@
+github.com/go-opentype/fonts v0.8.0 h1:77i3VPIH90GbstzNb21mk+an4WvEOe2idC6W+J0n0fw=
+github.com/go-opentype/fonts v0.8.0/go.mod h1:C6yQL2apHItfEZ5hztpsHF0S5mlX/hklLlq/Z5fRG/g=
+github.com/go-opentype/opentype v0.9.0 h1:GFgcJ3nwTDp4NJr5O+Paw7lhZx5Jv/R+noZwvhYDlkM=
+github.com/go-opentype/opentype v0.9.0/go.mod h1:AOixevJf7XQaH7+WG+OMIOZEbYPXfMqklVk26Y6YTUU=
+github.com/go-pdfkit/extract v0.1.0 h1:7IJDJMH43l2wJZD+DkW5StmeAS1NDt2FnT2LBceXTog=
+github.com/go-pdfkit/extract v0.1.0/go.mod h1:3ejdi87IrdR60wmjzLswdqzBzHCgFlrLtxkMxRN1KSU=
+github.com/go-pdfkit/pdffont v0.1.0 h1:mThWvPn3LLETup2adm9xMcNZWECP2Wq3eTSyTW4O3I0=
+github.com/go-pdfkit/pdffont v0.1.0/go.mod h1:y4vo5DgT95e57C3XxIWfA/xss+x6RwZwyj6KWdJc86s=
 github.com/go-pdfkit/reader v0.4.0 h1:qPbNZSO+Xl+4NBvQoV1PYt7HlqHaEAQ1tUc6/IL8JAU=
 github.com/go-pdfkit/reader v0.4.0/go.mod h1:fQFOVfCMUui1AdvD4qhimdyvvNr9KvvJ1S7IuKZjyV8=