diff --git a/packages/toolchain/math/mpfr/build b/packages/toolchain/math/mpfr/build
index 0450ef8817..dfdb754d36 100755
--- a/packages/toolchain/math/mpfr/build
+++ b/packages/toolchain/math/mpfr/build
@@ -3,7 +3,6 @@
 . config/options $1
 
 $SCRIPTS/install ccache
-$SCRIPTS/build gmp
 
 setup_toolchain host
 
diff --git a/packages/toolchain/math/mpfr/meta b/packages/toolchain/math/mpfr/meta
new file mode 100644
index 0000000000..9af9a290d5
--- /dev/null
+++ b/packages/toolchain/math/mpfr/meta
@@ -0,0 +1,14 @@
+PKG_NAME="mpfr"
+PKG_VERSION="3.0.0"
+PKG_REV="1"
+PKG_ARCH="any"
+PKG_LICENSE="LGPL"
+PKG_SITE="http://www.mpfr.org/"
+PKG_URL="http://www.mpfr.org/mpfr-current/$PKG_NAME-$PKG_VERSION.tar.bz2"
+PKG_DEPENDS=""
+PKG_BUILD_DEPENDS="gmp"
+PKG_PRIORITY="optional"
+PKG_SECTION="toolchain/math"
+PKG_SHORTDESC="mpfr: A C library for multiple-precision floating-point computations with exact roundi"
+PKG_LONGDESC="The MPFR library is a C library for multiple-precision floating-point computations with exact rounding (also called correct rounding). It is based on the GMP multiple-precision library. The main goal of MPFR is to provide a library for multiple-precision floating-point computation which is both efficient and has well-defined semantics. It copies the good ideas from the ANSI/IEEE-754 standard for double-precision floating-point arithmetic (53-bit mantissa)."
+PKG_IS_ADDON="no"
diff --git a/packages/toolchain/math/mpfr/url b/packages/toolchain/math/mpfr/url
deleted file mode 100644
index 8d8a1e0d76..0000000000
--- a/packages/toolchain/math/mpfr/url
+++ /dev/null
@@ -1 +0,0 @@
-http://www.mpfr.org/mpfr-current/mpfr-3.0.0.tar.bz2
\ No newline at end of file