diff --git a/toml.go b/toml.go
new file mode 100644
index 0000000..115d073
--- /dev/null
+++ b/toml.go
@@ -0,0 +1,467 @@
+package toml
+
+import (
+	"fmt"
+	"unicode/utf8"
+)
+
+func Unmarshal(data []byte, v interface{}) error {
+	// TODO
+	return nil
+}
+
+func Marshal(v interface{}) ([]byte, error) {
+	// TODO
+	return nil, nil
+}
+
+type Document struct {
+}
+
+type builder interface {
+	Whitespace(b []byte)
+	Comment(b []byte)
+	UnquotedKey(b []byte)
+	LiteralString(b []byte)
+	BasicString(b []byte)
+}
+
+type position struct {
+	line   int
+	column int
+}
+
+type documentBuilder struct {
+	document Document
+}
+
+func (d *documentBuilder) BasicString(b []byte) {
+	s := string(b)
+	fmt.Printf("BasicString: '%s'\n", s)
+}
+
+func (d *documentBuilder) LiteralString(b []byte) {
+	s := string(b)
+	fmt.Printf("LiteralString: '%s'\n", s)
+}
+
+func (d *documentBuilder) UnquotedKey(b []byte) {
+	s := string(b)
+	fmt.Printf("UnquotedKey: '%s'\n", s)
+}
+
+func (d *documentBuilder) Comment(b []byte) {
+	s := string(b)
+	fmt.Printf("Comment: '%s'\n", s)
+}
+
+func (d *documentBuilder) Whitespace(b []byte) {
+	s := string(b)
+	fmt.Printf("Whitespace: '%s'\n", s)
+}
+
+func Parse(b []byte) (Document, error) {
+	builder := documentBuilder{}
+	p := parser{builder: &builder}
+	err := p.parse(b)
+	if err != nil {
+		return Document{}, err
+	}
+	return builder.document, nil
+}
+
+// eof is a rune value indicating end-of-file.
+const eof = -1
+
+type parser struct {
+	builder builder
+}
+
+type InvalidCharacter struct {
+	r rune
+}
+
+func (e *InvalidCharacter) Error() string {
+	return fmt.Sprintf("unexpected character '%#U'", e.r)
+}
+
+type UnexpectedCharacter struct {
+	r        rune
+	expected rune
+}
+
+func (e *UnexpectedCharacter) Error() string {
+	return fmt.Sprintf("expected character '%#U' but got '%#U'", e.expected, e.r)
+}
+
+func (p *parser) parse(b []byte) error {
+	next := b
+	var err error
+	for {
+		next, err = p.parseExpression(next)
+		if err != nil {
+			return err
+		}
+		if len(next) == 0 {
+			return nil
+		}
+
+		// new lines between expressions
+		r, size, err := readRune(next)
+		if err != nil {
+			return err
+		}
+		if r == '\n' {
+			next = next[size:]
+			continue
+		}
+		if r == '\r' {
+			r, size2, err := readRune(next)
+			if err != nil {
+				return err
+			}
+			if r == '\n' {
+				next = next[size+size2:]
+				continue
+			}
+		}
+		return &InvalidCharacter{r: r}
+	}
+}
+
+func (p *parser) parseExpression(b []byte) ([]byte, error) {
+	next, err := p.parseWhitespace(b)
+	if err != nil {
+		return nil, err
+	}
+	r, _, err := readRune(next)
+	if err != nil {
+		return nil, err
+	}
+	// Line with just whitespace and a comment. We can exit early.
+	if r == '#' {
+		return p.parseComment(next)
+	}
+
+	// or line with something?
+
+	if r == '[' {
+		// parse table. could be either a standard table or an array table
+	}
+
+	// it has to be a keyval
+
+	if isUnquotedKeyRune(r) || r == '\'' || r == '"' {
+		next, err = p.parseKeyval(next)
+		if err != nil {
+			return nil, err
+		}
+	}
+
+	// parse trailing whitespace and comment
+
+	next, err = p.parseWhitespace(next)
+	if err != nil {
+		return nil, err
+	}
+
+	r, _, err = readRune(next)
+	if err != nil {
+		return nil, err
+	}
+
+	if r == '#' {
+		return p.parseComment(next)
+	}
+
+	return next, nil
+}
+
+func (p *parser) parseKeyval(b []byte) ([]byte, error) {
+	// key keyval-sep val
+	next, err := p.parseKey(b)
+	if err != nil {
+		return nil, err
+	}
+	return next, nil
+}
+
+func (p *parser) parseKey(b []byte) ([]byte, error) {
+	// simple-key / dotted-key
+	// dotted-key = simple-key 1*( dot-sep simple-key )
+
+	return p.parseSimpleKey(b)
+	// TODO: dotted key
+}
+
+func isUnquotedKeyRune(r rune) bool {
+	return (r >= 'A' && r <= 'Z') || (r >= 'a' && r <= 'z') || (r >= '0' && r <= '9') || r == '-' || r == '_'
+}
+
+func (p *parser) parseSimpleKey(b []byte) ([]byte, error) {
+	// simple-key = quoted-key / unquoted-key
+	// quoted-key = basic-string / literal-string
+	// unquoted-key = 1*( ALPHA / DIGIT / %x2D / %x5F ) ; A-Z / a-z / 0-9 / - / _
+	// basic-string = quotation-mark *basic-char quotation-mark
+	// literal-string = apostrophe *literal-char apostrophe
+
+	r, _, err := readRune(b)
+	if err != nil {
+		return nil, err
+	}
+
+	if r == '\'' {
+		return p.parseLiteralString(b)
+	}
+	if r == '"' {
+		return p.parseBasicString(b)
+	}
+
+	return p.parseUnquotedKey(b)
+}
+
+func (p *parser) parseUnquotedKey(b []byte) ([]byte, error) {
+	length := 0
+	r, size, err := readRune(b)
+	if err != nil {
+		return nil, err
+	}
+
+	if !isUnquotedKeyRune(r) {
+		return nil, &InvalidCharacter{r: r}
+	}
+	length += size
+
+	for {
+		r, size, err := readRune(b[length:])
+		if err != nil {
+			return nil, err
+		}
+		if !isUnquotedKeyRune(r) {
+			break
+		}
+		length += size
+	}
+	p.builder.UnquotedKey(b[:length])
+	return b[length:], nil
+}
+
+func expectRune(b []byte, expected rune) (int, error) {
+	r, size, err := readRune(b)
+	if err != nil {
+		return 0, err
+	}
+	if r != expected {
+		return 0, &UnexpectedCharacter{
+			r:        r,
+			expected: expected,
+		}
+	}
+	return size, nil
+}
+
+func (p *parser) parseComment(b []byte) ([]byte, error) {
+	length := 0
+
+	size, err := expectRune(b, '#')
+	if err != nil {
+		return b, err
+	}
+	length += size
+
+	for {
+		r, size, err := readRune(b[length:])
+		if err != nil {
+			return nil, err
+		}
+		if r == eof || r == '\n' {
+			if length > 0 {
+				p.builder.Comment(b[:length])
+			}
+			return b[length:], nil
+		}
+		length += size
+	}
+}
+
+func isWhitespace(r rune) bool {
+	switch r {
+	case 0x20, 0x09:
+		return true
+	default:
+		return false
+	}
+}
+
+type InvalidUnicodeError struct {
+	r rune
+}
+
+func (e *InvalidUnicodeError) Error() string {
+	return fmt.Sprintf("invalid unicode: %#U", e.r)
+}
+
+func readRune(b []byte) (rune, int, error) {
+	r, size := utf8.DecodeRune(b)
+	if r == utf8.RuneError {
+		if size == 0 { // eof
+			return eof, 0, nil
+		}
+		if size == 1 { // invalid rune
+			return utf8.RuneError, 1, &InvalidUnicodeError{r: r}
+		}
+	}
+	return r, size, nil
+}
+
+func (p *parser) parseWhitespace(b []byte) ([]byte, error) {
+	length := 0
+	for {
+		r, size, err := readRune(b[length:])
+		if err != nil {
+			return nil, err
+		}
+		if isWhitespace(r) {
+			length += size
+		} else {
+			if length > 0 {
+				p.builder.Whitespace(b[:length])
+			}
+			return b[length:], nil
+		}
+	}
+}
+
+func isNonAsciiChar(r rune) bool {
+	return (r >= 0x80 && r <= 0xD7FF) || (r >= 0xE000 && r <= 0x10FFFF)
+}
+
+func isLiteralChar(r rune) bool {
+	return r == 0x09 || (r >= 0x20 && r <= 0x26) || (r >= 0x28 && r <= 0x7E) || isNonAsciiChar(r)
+}
+
+func (p *parser) parseLiteralString(b []byte) ([]byte, error) {
+	// literal-string = apostrophe *literal-char apostrophe
+	// literal-char = %x09 / %x20-26 / %x28-7E / non-ascii
+	// non-ascii = %x80-D7FF / %xE000-10FFFF
+
+	length := 0
+
+	start, err := expectRune(b, '\'')
+	if err != nil {
+		return nil, err
+	}
+
+	for {
+		r, size, err := readRune(b[start+length:])
+		if err != nil {
+			return nil, err
+		}
+		if r == '\'' {
+			p.builder.LiteralString(b[start : start+length])
+			return b[start+length+size:], nil
+		}
+		if !isLiteralChar(r) {
+			return nil, &InvalidCharacter{r: r}
+		}
+		length += size
+	}
+}
+
+func isBasicStringChar(r rune) bool {
+	return r == ' ' || r == 0x21 || r >= 0x23 && r <= 0x5B || r >= 0x5D && r <= 0x7E || isNonAsciiChar(r)
+}
+
+func isEscapeChar(r rune) bool {
+	return r == '"' || r == '\\' || r == 'b' || r == 'f' || r == 'n' || r == 'r' || r == 't'
+}
+
+func isHex(r rune) bool {
+	return (r >= '0' && r <= '9') || (r >= 'A' && r <= 'F')
+}
+
+func (p *parser) parseBasicString(b []byte) ([]byte, error) {
+	// basic-string = quotation-mark *basic-char quotation-mark
+	// basic-char = basic-unescaped / escaped
+	// basic-unescaped = wschar / %x21 / %x23-5B / %x5D-7E / non-ascii
+	// escaped = escape escape-seq-char
+	//escape = %x5C                   ; \
+	//escape-seq-char =  %x22         ; "    quotation mark  U+0022
+	//escape-seq-char =/ %x5C         ; \    reverse solidus U+005C
+	//escape-seq-char =/ %x62         ; b    backspace       U+0008
+	//escape-seq-char =/ %x66         ; f    form feed       U+000C
+	//escape-seq-char =/ %x6E         ; n    line feed       U+000A
+	//escape-seq-char =/ %x72         ; r    carriage return U+000D
+	//escape-seq-char =/ %x74         ; t    tab             U+0009
+	//escape-seq-char =/ %x75 4HEXDIG ; uXXXX                U+XXXX
+	//escape-seq-char =/ %x55 8HEXDIG ; UXXXXXXXX            U+XXXXXXXX
+	// HEXDIG = DIGIT / "A" / "B" / "C" / "D" / "E" / "F"
+	length := 0
+
+	start, err := expectRune(b, '"')
+	if err != nil {
+		return nil, err
+	}
+
+	for {
+		r, size, err := readRune(b[start+length:])
+		if err != nil {
+			return nil, err
+		}
+		if r == '"' {
+			p.builder.BasicString(b[start : start+length])
+			return b[start+length+size:], nil
+		}
+
+		if r == '\\' {
+			length += size
+			r, size, err := readRune(b[start+length:])
+			if err != nil {
+				return nil, err
+			}
+
+			if isEscapeChar(r) {
+				length += size
+				continue
+			}
+
+			if r == 'u' {
+				length += size
+				for i := 0; i < 4; i++ {
+					r, size, err := readRune(b[start+length:])
+					if err != nil {
+						return nil, err
+					}
+					if !isHex(r) {
+						return nil, &InvalidCharacter{r: r}
+					}
+					length += size
+				}
+				continue
+			}
+
+			if r == 'U' {
+				length += size
+				for i := 0; i < 8; i++ {
+					r, size, err := readRune(b[start+length:])
+					if err != nil {
+						return nil, err
+					}
+					if !isHex(r) {
+						return nil, &InvalidCharacter{r: r}
+					}
+					length += size
+				}
+				continue
+			}
+
+			return nil, &InvalidCharacter{r: r}
+
+		}
+
+		if isBasicStringChar(r) {
+			length += size
+			continue
+		}
+	}
+}
diff --git a/toml_test.go b/toml_test.go
new file mode 100644
index 0000000..1aae1eb
--- /dev/null
+++ b/toml_test.go
@@ -0,0 +1,34 @@
+package toml
+
+import (
+	"fmt"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestParse(t *testing.T) {
+
+	inputs := []string{
+		`     #foo`,
+		`#foo`,
+		`#`,
+		"\n\n\n",
+		"#one\n   # two   \n",
+		`a`,
+		`abc`,
+		`  abc   # foo`,
+		`'abc'`,
+		`"foo bar"`,
+		`"hello\tworld"`,
+		`"hello \u1234 foo"`,
+	}
+
+	for i, data := range inputs {
+		t.Run(fmt.Sprintf("example %d", i), func(t *testing.T) {
+			doc, err := Parse([]byte(data))
+			assert.NoError(t, err)
+			fmt.Println(doc)
+		})
+	}
+}