diff --git a/internal/adapters/input/cli.go b/internal/adapters/input/cli.go
new file mode 100644
index 0000000..de1ab6f
--- /dev/null
+++ b/internal/adapters/input/cli.go
@@ -0,0 +1 @@
+package input
diff --git a/internal/adapters/input/input_test.go b/internal/adapters/input/input_test.go
new file mode 100644
index 0000000..33a504d
--- /dev/null
+++ b/internal/adapters/input/input_test.go
@@ -0,0 +1,56 @@
+package input
+
+import (
+	core "anemon/internal/core"
+	"reflect"
+	"testing"
+)
+
+func TestParagraphe(t *testing.T) {
+    work_input := `
+# Back-End Intern
+## February 2024 -- August 2024
+### TechCorp
+#### Internship
+- Assisted in developing and optimizing a key business process for expanding into new markets, collaborating with various teams to ensure compliance and smooth integration.
+- Participated in the migration of core backend logic from a monolithic application to a microservice architecture, improving system performance and scalability.
+- Enhanced system monitoring and reliability by implementing tracing mechanisms and performance objectives.
+
+# Back-End Intern
+## February 2024 -- August 2024
+### TechCorp
+#### Internship
+- Assisted in developing and optimizing a key business process for expanding into new markets, collaborating with various teams to ensure compliance and smooth integration.
+- Participated in the migration of core backend logic from a monolithic application to a microservice architecture, improving system performance and scalability.
+- Enhanced system monitoring and reliability by implementing tracing mechanisms and performance objectives.`
+
+    work_paragraphe1 := core.Paragraphe{H1: "Back-End Intern", H2: "February 2024 -- August 2024",
+        H3: "TechCorp", H4: "Internship",Items: []string{
+            "Assisted in developing and optimizing a key business process for expanding into new markets, collaborating with various teams to ensure compliance and smooth integration.",
+            "Participated in the migration of core backend logic from a monolithic application to a microservice architecture, improving system performance and scalability.",
+            "Enhanced system monitoring and reliability by implementing tracing mechanisms and performance objectives."}}
+
+    invalid_input := "ajsdlhsaeld##dafdbhkbhkjsd##"
+    work_expected_result := []core.Paragraphe{work_paragraphe1,work_paragraphe1}
+
+    t.Run("Work Paragraphes should return a slice of valid Paragraphe", func (t *testing.T) {
+        result := getParagrapheFrom(work_input)
+
+        if !reflect.DeepEqual(result[0], work_paragraphe1){
+            t.Fatalf("the first Paragraphe should be :\n%s\n got :%s", work_expected_result[0].Items[0],result[0].Items[0])
+        }
+
+        if !reflect.DeepEqual(result[1], work_paragraphe1){
+            t.Fatalf("the second Paragraphe should be :\n%s\n got :%s", work_expected_result[0].Items[0],result[0].Items[0])
+        }
+    })
+
+    t.Run("Invalid inout should return nothing", func (t *testing.T) {
+        result := getParagrapheFrom(invalid_input)
+
+        if result != nil{
+            t.Fatalf("Invalid input should return nil got %s",result)
+        }
+    })
+
+}
diff --git a/internal/adapters/input/markdown_tree_reader.go b/internal/adapters/input/markdown_tree_reader.go
new file mode 100644
index 0000000..57e7b95
--- /dev/null
+++ b/internal/adapters/input/markdown_tree_reader.go
@@ -0,0 +1,130 @@
+package input
+
+import (
+	core "anemon/internal/core"
+	"errors"
+	"fmt"
+	"os"
+	"path/filepath"
+	"regexp"
+	"strings"
+)
+
+// `GetCVFrom` takes a root directory and extracts Markdown documents to generate a list of CV type.
+//
+// This function assumes the directory structure is a tree with a depth of 3, where each leaf
+// is a Markdown (.md) document. Each document may contain multiple paragraphs, but the headers
+// should not repeat within the same document.
+func GetCVsFrom(root string) ([]core.CV, error) {
+    cvs := make([]core.CV,0)
+    current_lang := ""
+    current_sections := make([]core.Section,4)
+    //Walk
+    err := filepath.Walk(root, func(path string, info os.FileInfo, err error) error {
+        if err != nil {
+            return err
+        }
+        if info.IsDir(){
+            current_lang = info.Name()
+            cvs = append(cvs, core.CV{Lang: current_lang}) 
+        }
+        if !info.IsDir() && strings.HasSuffix(info.Name(), ".md") {
+            if current_lang == "" {
+                return errors.New("markdown file found before lang directory")
+            }
+            content, err := os.ReadFile(path)
+            if err != nil {
+                return err
+            }
+            new_section := core.Section{Title:info.Name(), Paragraphes: getParagrapheFrom(string(content))}
+            current_sections = append(current_sections, new_section)
+            if info.Name()=="skill.md"{//Walk is in alphabetique order, therefore skill is last
+                cvs[len(cvs)-1].Sections=current_sections
+            }
+        }
+        return nil
+    })
+    if err != nil {
+        return nil, err
+    }
+
+    return cvs,nil
+}
+
+
+//Take string in the md format and return a slice of core.Paragraphe for each paragraph inside of it
+//We define a paragraph by text block where \n\n indicate a separation
+func getParagrapheFrom(s_section string)[]core.Paragraphe{
+    paragraphs := make([]core.Paragraphe,0)
+    for _,paragraphe := range strings.Split(s_section, "\n\n"){
+        n_paragraphe,err := parse_paragraphe(paragraphe)
+        if is_empty(n_paragraphe){
+            continue
+        }
+        if err != nil {
+            fmt.Println("Failed to parse paragraphe ") 
+        } else{
+            paragraphs = append(paragraphs, n_paragraphe)
+        }
+    }
+    if len(paragraphs) == 0 { return nil }
+    return paragraphs
+}
+
+func is_empty(p core.Paragraphe)bool{
+    no_header := p.H1 == "" && p.H2 == "" &&  p.H3 == "" && p.H4 == ""
+    no_items := len(p.Items)==0
+    return no_header && no_items
+}
+
+/*
+Parse parses a Markdown-like `paragraph` into a `Paragraph`,
+extracting headings and description based on the number of leading hashtags or stars.
+Returns an error if the format is invalid.
+*/
+func parse_paragraphe(paragraph string) (core.Paragraphe,error){
+
+    n_paragraphe := core.Paragraphe{}
+    if len(strings.Split(paragraph, "\n\n")) > 1{
+        return n_paragraphe, errors.New("Tried to parse mutiple paragraph into a single section")
+    }
+    hashtag_regex, _ := regexp.Compile("^#+")
+    wasASkill := false
+    for _, line := range strings.Split(strings.TrimRight(paragraph, "\n"), "\n") {
+        nb_hashtag := len(hashtag_regex.FindString(line))
+
+        if len(line) == 0{ continue }
+
+        if wasASkill {
+            wasASkill = false
+            n_paragraphe.H2= strings.TrimLeft(line,"- ")
+        }
+
+        if nb_hashtag == 0 && string(line[0])=="*" && len(strings.Trim(line,"*")) == len(line) - 4 {//Trim should **tt** -> tt
+            n_paragraphe.H1= strings.Trim(line,"*")
+            wasASkill = true
+        }
+
+
+        switch{
+        case nb_hashtag>0 && line[nb_hashtag] != ' ':
+            return n_paragraphe, errors.New("Err: cannot parse this md line{"+line+"}  # should be followed by space")
+        case nb_hashtag == 1:
+            n_paragraphe.H1=line[nb_hashtag+1:]
+        case nb_hashtag == 2:
+            n_paragraphe.H2=line[nb_hashtag+1:]
+        case nb_hashtag == 3:
+            n_paragraphe.H3=line[nb_hashtag+1:]
+        case nb_hashtag == 4:
+            n_paragraphe.H4=line[nb_hashtag+1:]
+        case nb_hashtag == 0 && len(line)>1:
+            if line[:2] == "- "{
+                n_paragraphe.Items = append(n_paragraphe.Items, strings.TrimLeft(line,"- "))
+            }
+        case nb_hashtag > 4:
+            return n_paragraphe, errors.New("Err: cannot parse this md line{"+line+"}")
+    }
+    }
+    return n_paragraphe, nil
+}
+
diff --git a/internal/core/ports.go b/internal/core/ports.go
new file mode 100644
index 0000000..68469b0
--- /dev/null
+++ b/internal/core/ports.go
@@ -0,0 +1,20 @@
+package core
+
+//Test
+type CV struct {
+    Lang string
+    Sections []Section
+}
+
+type Section struct {
+    Title string
+    Paragraphes []Paragraphe
+}
+	
+type Paragraphe struct {
+    H1 string
+    H2  string
+    H3  string
+    H4  string
+    Items  []string
+}
diff --git a/internal/walker/cv.go b/internal/walker/cv.go
index ffd93a4..315d044 100644
--- a/internal/walker/cv.go
+++ b/internal/walker/cv.go
@@ -12,9 +12,8 @@ import (
 func WalkCV(root string) (map[string]map[string]string, error) {
 	fileMap := make(map[string]string)
 	err := filepath.Walk(root, func(path string, info os.FileInfo, err error) error {
-		if err != nil {
-			return err
-		}
+                println(info.Name())
+                println(info.IsDir())
 		if !info.IsDir() && strings.HasSuffix(info.Name(), ".md") {
 			relativePath := strings.TrimSuffix(strings.TrimPrefix(path, root+string(os.PathSeparator)), ".md")
 			content, err := os.ReadFile(path)