diff --git a/axl_string.c b/axl_string.c
index 6707b3a..db7b3f9 100644
--- a/axl_string.c
+++ b/axl_string.c
@@ -263,3 +263,171 @@ i8* axl_itoa(i64 val, i8* str, u8 base)
 
     return axl_strrev(str);
 }
+
+#if !__STDC_HOSTED__
+int _fltused;
+#endif // !__STDC_HOSTED__
+
+/*
+ * @brief Converts float to string with specified precision
+ * 
+ * @param val: Float value to convert
+ * @param str: Destination buffer (must be large enough)
+ * @param precision: Digits after decimal point (0-AXL_F32_MAX_PRECISION)
+ * @return Pointer to start of string
+ * 
+ * @note Handles nan, inf, -inf, zero
+ * @note Uses scientific notation for very large/small numbers
+ * @note Removes trailing zeros from fractional part
+ */
+i8* axl_ftoa(f32 val, i8* str, u32 precision)
+{
+    if (!str) return NULL;
+    
+    // Clamp precision to meaningful range
+    if (precision > AXL_F32_MAX_PRECISION) 
+    {
+        precision = AXL_F32_MAX_PRECISION;
+    }
+    
+    // Extract float components
+    union { f32 f; u32 u; } bits = { .f = val };
+    u32 sign = bits.u >> 31;
+    u32 exponent = (bits.u >> 23) & 0xFF;
+    u32 mantissa = bits.u & 0x7FFFFF;
+    
+    // Handle NaN
+    if (exponent == 0xFF && mantissa != 0)
+	{
+        axl_strcpy(str, "nan");
+        return str;
+    }
+    
+    // Handle Infinity
+    if (exponent == 0xFF && mantissa == 0)
+	{
+        if (sign) axl_strcpy(str, "-inf");
+        else axl_strcpy(str, "inf");
+        return str;
+    }
+    
+    // Handle zero and denormalized numbers (treat as zero)
+    if (exponent == 0)
+	{
+        i8* p = str;
+        if (sign) *p++ = '-';
+        *p++ = '0';
+        if (precision > 0)
+        {
+            *p++ = '.';
+            for (u32 i = 0; i < precision; i++)
+            {
+                *p++ = '0';
+            }
+        }
+        *p = '\0';
+        return str;
+    }
+    
+    i8* start = str;
+    f32 abs_val = val;
+    
+    // Apply sign
+    if (sign)
+	{
+        *str++ = '-';
+        abs_val = -abs_val;
+    }
+    
+    // Calculate magnitude for format selection
+    i32 magnitude = 0;
+    f32 temp = abs_val;
+    while (temp >= 10.0f)
+	{ temp /= 10.0f; magnitude++; }
+    while (temp > 0.0f && temp < 1.0f)
+	{ temp *= 10.0f; magnitude--; }
+    
+    // Use scientific notation for extreme ranges
+    if (magnitude > 15 || magnitude < -15)
+	{
+        // Normalize mantissa to [1, 10)
+        f32 mant = abs_val;
+        while (mant >= 10.0f) mant /= 10.0f;
+        while (mant > 0.0f && mant < 1.0f) mant *= 10.0f;
+        
+        // Integer part (single digit)
+        u64 int_part = (u64)mant;
+        axl_utoa(int_part, str, 10);
+        str += axl_strlen(str);
+        
+        // Fractional part
+        if (precision > 0)
+        {
+            *str++ = '.';
+            f32 frac = mant - (f32)int_part;
+            for (u32 i = 0; i < precision; i++)
+            {
+                frac *= 10.0f;
+                u32 digit = (u32)frac;
+                *str++ = '0' + digit;
+                frac -= (f32)digit;
+            }
+            // Trim trailing zeros
+            while (*(str-1) == '0' && *(str-2) != '.') str--;
+        }
+        
+        // Add exponent
+        *str++ = 'e';
+        if (magnitude >= 0) *str++ = '+';
+        axl_itoa(magnitude, str, 10);
+        return start;
+    }
+    
+    // Fixed-point representation for normal range
+    
+    // Split into integer and fractional parts
+    u64 int_part = (u64)abs_val;
+    f32 frac_part = abs_val - (f32)int_part;
+    
+    // Write integer part
+    if (int_part == 0)
+	{
+        *str++ = '0';
+    } 
+    else 
+    {
+        i8 buf[32];
+        axl_utoa(int_part, buf, 10);
+        axl_strcpy(str, buf);
+        str += axl_strlen(buf);
+    }
+    
+    // Write fractional part
+    if (precision > 0)
+	{
+        *str++ = '.';
+        f32 f = frac_part;
+        u32 i;
+        for (i = 0; i < precision; i++)
+	    {
+            f *= 10.0f;
+            u32 digit = (u32)f;
+            *str++ = '0' + digit;
+            f -= (f32)digit;
+        }
+        
+        // Remove trailing zeros
+        while (str > start + 2 && *(str-1) == '0')
+	    {
+            str--;
+        }
+        // Remove decimal point if no fractional digits remain
+        if (*(str-1) == '.')
+	    {
+            str--;
+        }
+    }
+    
+    *str = '\0';
+    return start;
+}
diff --git a/axl_string.h b/axl_string.h
index ffd6878..22afe61 100644
--- a/axl_string.h
+++ b/axl_string.h
@@ -14,5 +14,6 @@ i8*       axl_strstr(const i8* str, const i8* substr);
 i8*       axl_strrev(i8* str);
 i8*       axl_utoa(u64 val, i8* str, u8 base);
 i8*       axl_itoa(i64 val, i8* str, u8 base);
+i8*       axl_ftoa(f32 val, i8* str, u32 precision);
 
 #endif // AXL_STRING
diff --git a/axl_types.h b/axl_types.h
index 423f797..6876a57 100644
--- a/axl_types.h
+++ b/axl_types.h
@@ -8,6 +8,7 @@ typedef signed short        i16;
 typedef unsigned short      u16;
 typedef signed int          i32;
 typedef unsigned int        u32;
+typedef float               f32;
 typedef signed long long    i64;
 typedef unsigned long long  u64;
 
@@ -32,4 +33,6 @@ typedef unsigned long long  u64;
 #define I64_MAX  9223372036854775807LL
 #define U64_MAX  18446744073709551615ULL
 
+#define AXL_F32_MAX_PRECISION 6
+
 #endif // !AXL_TYPES_H
diff --git a/main.c b/main.c
index a601f20..df39e82 100644
--- a/main.c
+++ b/main.c
@@ -4,6 +4,7 @@
 #include "axl_string.h"
 #include "axl_rle.h"
 #include "axl_memory.h"
+#include "axl_types.h"
 #include <windows.h>
 
 void* memset(void *s, int c, size_t n)
@@ -89,6 +90,12 @@ int _start(void)
     axl_strcat(text_buff, enc_size_str);
     axl_puts(text_buff);
 
+
+
+    axl_memset(text_buff, '\0', sizeof(text_buff));
+    axl_ftoa(666.32456f, text_buff, 2);
+    axl_puts(text_buff);
+
     /*u32 new_len = 0;*/
     /*u8 src[] = "Bonjour le monde!";*/
     /*u32 src_len = axl_strlen((i8*)src);*/
diff --git a/makefile b/makefile
index 661ef4c..211dbda 100644
--- a/makefile
+++ b/makefile
@@ -1,29 +1,40 @@
 CC = clang
-AXL_SOURCES = $(filter-out main.c %_test.c, $(wildcard *.c))
-TEST_EXES = $(patsubst %.c,%.exe,$(wildcard *_test.c))
-TEST_CFLAGS = -Wall -Wextra -Werror -pedantic -std=c17 -static -Oz -fsigned-char
-CFLAGS = $(TEST_CFLAGS) -nostdlib -ffreestanding -fno-builtin -mno-stack-arg-probe
-LDFLAGS = -Wl,/SUBSYSTEM:CONSOLE,/ENTRY:_start -fuse-ld=lld
-LIBS = -lkernel32
-SOURCES = $(AXL_SOURCES) main.c
-TARGET = prog.exe
 
-all: clean test $(TARGET)
+AXL_SOURCES = $(filter-out main.c %_test.c, $(wildcard *.c))
+TEST_EXES   = $(patsubst %.c,%.exe,$(wildcard *_test.c))
+
+TEST_CFLAGS = -Wall -Wextra -Werror -pedantic -std=c17 -static -Oz -fsigned-char
+
+MINIMAL_CFLAGS = -nostdlib -ffreestanding -fno-builtin -mno-stack-arg-probe \
+                 -fno-stack-protector -fno-unwind-tables -fno-exceptions -fno-rtti \
+                 -fno-asynchronous-unwind-tables -fno-ident
+
+LDFLAGS = -fuse-ld=lld -static \
+          -Wl,/SUBSYSTEM:CONSOLE \
+          -Wl,/ENTRY:_start \
+          -Wl,/OPT:NOICF \
+          -Wl,/OPT:NOREF \
+          -Wl,/NODEFAULTLIB \
+          -Wl,/MERGE:.rdata=.text
+
+LIBS = -lkernel32
+
+SOURCES = $(AXL_SOURCES) main.c
+TARGET  = prog.exe
+
+all: clean $(TARGET)
 
 %_test.exe: %_test.c
 	$(CC) $(TEST_CFLAGS) $(AXL_SOURCES) $< -o $@
 
-test: $(TEST_EXES)
-	for %%i in ($(TEST_EXES)) do ( \
-		echo Running %%i && \
-		%%i || exit /b 1 \
-	)
+test: clean $(TEST_EXES)
+	@echo "=== Running tests ==="
+	@for %%i in ($(TEST_EXES)) do (echo ---- %%i ---- && %%i || exit /b 1)
 
-$(TARGET): $(SOURCES)
-	$(CC) $(CFLAGS) $(SOURCES) -o $(TARGET) $(LDFLAGS) $(LIBS)
+$(TARGET): $(SOURCES) 
+	$(CC) $(TEST_CFLAGS) $(MINIMAL_CFLAGS) $^ -o $@ $(LDFLAGS) $(LIBS)
 
 clean:
-	rm -f $(TARGET)
-	rm -f $(TEST_EXES)
+	-del /q *.exe *.o 2>nul 
 
 .PHONY: all clean test