diff --git a/markup/img.go b/markup/img.go
index 18e75d8..244451b 100644
--- a/markup/img.go
+++ b/markup/img.go
@@ -127,7 +127,7 @@ func (img Img) ToHtml() (html string) {
 					if i > 0 {
 						html += `<br>`
 					}
-					html += ParagraphToHtml(line)
+					html += ParagraphToHtml(img.hyphaName, line)
 				}
 			}
 			html += `</figcaption>`
diff --git a/markup/img_test.go b/markup/img_test.go
index ea9b6f2..f2e2044 100644
--- a/markup/img_test.go
+++ b/markup/img_test.go
@@ -12,14 +12,14 @@ func TestParseStartOfEntry(t *testing.T) {
 		entry          imgEntry
 		followedByDesc bool
 	}{
-		{"apple", imgEntry{"apple", "", "", ""}, false},
-		{"pear:", imgEntry{"pear", "", "", ""}, false},
-		{"яблоко: 30*60", imgEntry{"яблоко", "30", "60", ""}, false},
-		{"груша   : 65  ", imgEntry{"груша", "65", "", ""}, false},
-		{"жеронимо : 30 { full desc }", imgEntry{"жеронимо", "30", "", " full desc "}, false},
-		{"жорно жованна :   *5555 {partial description", imgEntry{"жорно_жованна", "", "5555", "partial description"}, true},
-		{"иноске : {full}", imgEntry{"иноске", "", "", "full"}, false},
-		{"j:{partial", imgEntry{"j", "", "", "partial"}, true},
+		{"apple", imgEntry{"/binary/apple", "", "", ""}, false},
+		{"pear|", imgEntry{"/binary/pear", "", "", ""}, false},
+		{"яблоко| 30*60", imgEntry{"/binary/яблоко", "30", "60", ""}, false},
+		{"груша   | 65  ", imgEntry{"/binary/груша", "65", "", ""}, false},
+		{"жеронимо | 30 { full desc }", imgEntry{"/binary/жеронимо", "30", "", " full desc "}, false},
+		{"жорно жованна |   *5555 {partial description", imgEntry{"/binary/жорно_жованна", "", "5555", "partial description"}, true},
+		{"иноске | {full}", imgEntry{"/binary/иноске", "", "", "full"}, false},
+		{"j|{partial", imgEntry{"/binary/j", "", "", "partial"}, true},
 	}
 	for _, triplet := range tests {
 		entry, followedByDesc := img.parseStartOfEntry(triplet.line)
diff --git a/markup/lexer.go b/markup/lexer.go
index b29b7fc..cad18d7 100644
--- a/markup/lexer.go
+++ b/markup/lexer.go
@@ -3,7 +3,6 @@ package markup
 import (
 	"fmt"
 	"html"
-	"path"
 	"strings"
 )
 
@@ -31,50 +30,6 @@ type Line struct {
 	contents interface{}
 }
 
-// Parse markup line starting with "=>" according to wikilink rules.
-// See http://localhost:1737/page/wikilink
-func wikilink(src string, state *GemLexerState) (href, text, class string) {
-	src = strings.TrimSpace(remover("=>")(src))
-	if src == "" {
-		return
-	}
-	// Href is text after => till first whitespace
-	href = strings.Fields(src)[0]
-	// Text is everything after whitespace.
-	// If there's no text, make it same as href
-	if text = strings.TrimPrefix(src, href); text == "" {
-		text = href
-	}
-
-	class = "wikilink_internal"
-
-	switch {
-	case strings.HasPrefix(href, "./"):
-		hyphaName := canonicalName(path.Join(
-			state.name, strings.TrimPrefix(href, "./")))
-		if !HyphaExists(hyphaName) {
-			class += " wikilink_new"
-		}
-		href = path.Join("/page", hyphaName)
-	case strings.HasPrefix(href, "../"):
-		hyphaName := canonicalName(path.Join(
-			path.Dir(state.name), strings.TrimPrefix(href, "../")))
-		if !HyphaExists(hyphaName) {
-			class += " wikilink_new"
-		}
-		href = path.Join("/page", hyphaName)
-	case strings.HasPrefix(href, "/"):
-	case strings.ContainsRune(href, ':'):
-		class = "wikilink_external"
-	default:
-		if !HyphaExists(canonicalName(href)) {
-			class += " wikilink_new"
-		}
-		href = path.Join("/page", href)
-	}
-	return href, strings.TrimSpace(text), class
-}
-
 func lex(name, content string) (ast []Line) {
 	var state = GemLexerState{name: name}
 
@@ -141,7 +96,7 @@ preformattedState:
 listState:
 	switch {
 	case startsWith("* "):
-		state.buf += fmt.Sprintf("\t<li>%s</li>\n", ParagraphToHtml(line[2:]))
+		state.buf += fmt.Sprintf("\t<li>%s</li>\n", ParagraphToHtml(state.name, line[2:]))
 	case startsWith("```"):
 		state.where = "pre"
 		addLine(state.buf + "</ul>")
@@ -157,7 +112,7 @@ listState:
 numberState:
 	switch {
 	case startsWith("*. "):
-		state.buf += fmt.Sprintf("\t<li>%s</li>\n", ParagraphToHtml(line[3:]))
+		state.buf += fmt.Sprintf("\t<li>%s</li>\n", ParagraphToHtml(state.name, line[3:]))
 	case startsWith("```"):
 		state.where = "pre"
 		addLine(state.buf + "</ol>")
@@ -209,9 +164,9 @@ normalState:
 		addLine(fmt.Sprintf(
 			"<blockquote id='%d'>%s</blockquote>", state.id, remover(">")(line)))
 	case startsWith("=>"):
-		source, content, class := wikilink(line, state)
+		href, text, class := Rocketlink(line, state.name)
 		addLine(fmt.Sprintf(
-			`<p><a id='%d' class='%s' href="%s">%s</a></p>`, state.id, class, source, content))
+			`<p><a id='%d' class='rocketlink %s' href="%s">%s</a></p>`, state.id, class, href, text))
 
 	case startsWith("<="):
 		addLine(parseTransclusion(line, state.name))
@@ -221,6 +176,6 @@ normalState:
 		state.where = "img"
 		state.img = ImgFromFirstLine(line, state.name)
 	default:
-		addLine(fmt.Sprintf("<p id='%d'>%s</p>", state.id, ParagraphToHtml(line)))
+		addLine(fmt.Sprintf("<p id='%d'>%s</p>", state.id, ParagraphToHtml(state.name, line)))
 	}
 }
diff --git a/markup/lexer_test.go b/markup/lexer_test.go
index 41dc64b..3298a36 100644
--- a/markup/lexer_test.go
+++ b/markup/lexer_test.go
@@ -41,13 +41,13 @@ func TestLex(t *testing.T) {
 </ul>`},
 		{6, "<p id='6'>text</p>"},
 		{7, "<p id='7'>more text</p>"},
-		{8, `<p><a id='8' class='wikilink_internal' href="/page/Pear">some link</a></p>`},
+		{8, `<p><a id='8' class='rocketlink wikilink_internal' href="/page/Pear">some link</a></p>`},
 		{9, `<ul id='9'>
 	<li>lin&#34;+</li>
 </ul>`},
 		{10, `<pre id='10' alt='alt text goes here' class='codeblock'><code>=&gt; preformatted text
 where markup is not lexed</code></pre>`},
-		{11, `<p><a id='11' class='wikilink_internal' href="/page/linking">linking</a></p>`},
+		{11, `<p><a id='11' class='rocketlink wikilink_internal' href="/page/linking">linking</a></p>`},
 		{12, "<p id='12'>text</p>"},
 		{13, `<pre id='13' alt='' class='codeblock'><code>()
 /\</code></pre>`},
@@ -56,11 +56,11 @@ where markup is not lexed</code></pre>`},
 			hyphaName: "Apple",
 			inDesc:    false,
 			entries: []imgEntry{
-				{"hypha1", "", "", ""},
-				{"hypha2", "", "", ""},
-				{"hypha3", "60", "", ""},
-				{"hypha4", "", "", " line1\nline2\n"},
-				{"hypha5", "", "", "\nstate of minnesota\n"},
+				{"/binary/hypha1", "", "", ""},
+				{"/binary/hypha2", "", "", ""},
+				{"/binary/hypha3", "60", "", ""},
+				{"/binary/hypha4", "", "", " line1\nline2\n"},
+				{"/binary/hypha5", "", "", "\nstate of minnesota\n"},
 			},
 		}},
 	})
diff --git a/markup/link.go b/markup/link.go
new file mode 100644
index 0000000..eb52c63
--- /dev/null
+++ b/markup/link.go
@@ -0,0 +1,49 @@
+package markup
+
+import (
+	"path"
+	"strings"
+)
+
+// LinkParts determines what href, text and class should resulting <a> have based on mycomarkup's addr, display and hypha name.
+//
+// => addr display
+// [[addr|display]]
+func LinkParts(addr, display, hyphaName string) (href, text, class string) {
+	if display == "" {
+		text = addr
+	} else {
+		text = strings.TrimSpace(display)
+	}
+	class = "wikilink_internal"
+
+	switch {
+	case strings.ContainsRune(addr, ':'):
+		return addr, text, "wikilink_external"
+	case strings.HasPrefix(addr, "/"):
+		return addr, text, class
+	case strings.HasPrefix(addr, "./"):
+		hyphaName = canonicalName(path.Join(hyphaName, addr[2:]))
+	case strings.HasPrefix(addr, "../"):
+		hyphaName = canonicalName(path.Join(path.Dir(hyphaName), addr[3:]))
+	default:
+		hyphaName = canonicalName(addr)
+	}
+	if !HyphaExists(hyphaName) {
+		class += " wikilink_new"
+	}
+	return "/page/" + hyphaName, text, class
+}
+
+// Parse markup line starting with "=>" according to wikilink rules.
+// See http://localhost:1737/page/wikilink
+func Rocketlink(src, hyphaName string) (href, text, class string) {
+	src = strings.TrimSpace(src[2:]) // Drop =>
+	if src == "" {
+		return
+	}
+	// Href is text after => till first whitespace
+	addr := strings.Fields(src)[0]
+	display := strings.TrimPrefix(src, addr)
+	return LinkParts(addr, display, hyphaName)
+}
diff --git a/markup/paragraph.go b/markup/paragraph.go
index 75070a4..1eb1f9b 100644
--- a/markup/paragraph.go
+++ b/markup/paragraph.go
@@ -17,6 +17,7 @@ const (
 	spanSuper
 	spanSub
 	spanMark
+	spanLink
 )
 
 func tagFromState(stt spanTokenType, tagState map[spanTokenType]bool, tagName, originalForm string) string {
@@ -32,7 +33,33 @@ func tagFromState(stt spanTokenType, tagState map[spanTokenType]bool, tagName, o
 	}
 }
 
-// getTextNode splits the `p` into two parts `textNode` and `rest` by the first encountered rune that resembles a span tag. If there is none, `textNode = p`, `rest = ""`. It handles escaping with backslash.
+func getLinkNode(input *bytes.Buffer, hyphaName string) string {
+	input.Next(2)
+	var (
+		escaping   = false
+		addrBuf    = bytes.Buffer{}
+		displayBuf = bytes.Buffer{}
+		currBuf    = &addrBuf
+	)
+	for input.Len() != 0 {
+		b, _ := input.ReadByte()
+		if escaping {
+			currBuf.WriteByte(b)
+			escaping = false
+		} else if b == '|' && currBuf == &addrBuf {
+			currBuf = &displayBuf
+		} else if b == ']' && bytes.HasPrefix(input.Bytes(), []byte{']'}) {
+			input.Next(1)
+			break
+		} else {
+			currBuf.WriteByte(b)
+		}
+	}
+	href, text, class := LinkParts(addrBuf.String(), displayBuf.String(), hyphaName)
+	return fmt.Sprintf(`<a href="%s" class="%s">%s</a>`, href, class, html.EscapeString(text))
+}
+
+// getTextNode splits the `input` into two parts `textNode` and `rest` by the first encountered rune that resembles a span tag. If there is none, `textNode = input`, `rest = ""`. It handles escaping with backslash.
 func getTextNode(input *bytes.Buffer) string {
 	var (
 		textNodeBuffer = bytes.Buffer{}
@@ -51,7 +78,7 @@ func getTextNode(input *bytes.Buffer) string {
 			escaping = false
 		} else if b == '\\' {
 			escaping = true
-		} else if strings.IndexByte("/*`^,!", b) >= 0 {
+		} else if strings.IndexByte("/*`^,![", b) >= 0 {
 			input.UnreadByte()
 			break
 		} else {
@@ -61,7 +88,7 @@ func getTextNode(input *bytes.Buffer) string {
 	return textNodeBuffer.String()
 }
 
-func ParagraphToHtml(input string) string {
+func ParagraphToHtml(hyphaName, input string) string {
 	var (
 		p   = bytes.NewBufferString(input)
 		ret strings.Builder
@@ -73,6 +100,7 @@ func ParagraphToHtml(input string) string {
 			spanSuper:  false,
 			spanSub:    false,
 			spanMark:   false,
+			spanLink:   false,
 		}
 		startsWith = func(t string) bool {
 			return bytes.HasPrefix(p.Bytes(), []byte(t))
@@ -99,6 +127,8 @@ func ParagraphToHtml(input string) string {
 		case startsWith("!!"):
 			ret.WriteString(tagFromState(spanMark, tagState, "mark", "!!"))
 			p.Next(2)
+		case startsWith("[["):
+			ret.WriteString(getLinkNode(p, hyphaName))
 		default:
 			ret.WriteString(html.EscapeString(getTextNode(p)))
 		}
@@ -119,6 +149,8 @@ func ParagraphToHtml(input string) string {
 				ret.WriteString(tagFromState(spanSub, tagState, "sub", ",,"))
 			case spanMark:
 				ret.WriteString(tagFromState(spanMark, tagState, "mark", "!!"))
+			case spanLink:
+				ret.WriteString(tagFromState(spanLink, tagState, "a", "[["))
 			}
 		}
 	}
diff --git a/markup/paragraph_test.go b/markup/paragraph_test.go
index 49505a9..6fd1d91 100644
--- a/markup/paragraph_test.go
+++ b/markup/paragraph_test.go
@@ -35,10 +35,15 @@ func TestParagraphToHtml(t *testing.T) {
 		{"this is a left **bold", "this is a left <strong>bold</strong>"},
 		{"this line has a ,comma, two of them", "this line has a ,comma, two of them"},
 		{"Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.", "Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum."},
+		{"A [[simple]] link", `A <a href="/page/simple" class="wikilink_internal">simple</a> link`},
 	}
 	for _, test := range tests {
-		if ParagraphToHtml(test[0]) != test[1] {
-			t.Error(fmt.Sprintf("%q: Wanted %q, got %q", test[0], test[1], ParagraphToHtml(test[0])))
+		if ParagraphToHtml("Apple", test[0]) != test[1] {
+			t.Error(fmt.Sprintf("%q: Wanted %q, got %q", test[0], test[1], ParagraphToHtml("Apple", test[0])))
 		}
 	}
 }
+
+func init() {
+	HyphaExists = func(_ string) bool { return true }
+}
diff --git a/markup/testdata/test.myco b/markup/testdata/test.myco
index 389188e..fd5880f 100644
--- a/markup/testdata/test.myco
+++ b/markup/testdata/test.myco
@@ -25,13 +25,13 @@ text
 
 img {
 hypha1
-hypha2:
-hypha3: 60
-hypha4: { line1
+hypha2|
+hypha3| 60
+hypha4| { line1
 line2
 } this is ignored
 
-hypha5: {
+hypha5| {
 state of minnesota
 }
 }
diff --git a/metarrhiza b/metarrhiza
index 9fa5334..d78a907 160000
--- a/metarrhiza
+++ b/metarrhiza
@@ -1 +1 @@
-Subproject commit 9fa5334ee958c02fbb327c002bf646e648a5a1be
+Subproject commit d78a90718ae9c36f7a114901ddad84b0e23221b3