diff --git a/zokrates_cli/src/bin.rs b/zokrates_cli/src/bin.rs
index 40153d08..c1ab68dc 100644
--- a/zokrates_cli/src/bin.rs
+++ b/zokrates_cli/src/bin.rs
@@ -380,21 +380,11 @@ fn main() {
                 })
                 .collect();
 
-            let witness_map = program_ast
-                .execute(arguments)
+            let witness = program_ast
+                .execute(&arguments)
                 .unwrap_or_else(|e| panic!(format!("Execution failed: {}", e)));
 
-            println!(
-                "\nWitness: \n\n{}",
-                witness_map
-                    .iter()
-                    .filter_map(|(variable, value)| match variable {
-                        variable if variable.is_output() => Some(format!("{} {}", variable, value)),
-                        _ => None,
-                    })
-                    .collect::<Vec<String>>()
-                    .join("\n")
-            );
+            println!("\nWitness: \n\n{}", witness.format_outputs());
 
             // write witness to file
             let output_path = Path::new(sub_matches.value_of("output").unwrap());
@@ -403,10 +393,7 @@ fn main() {
                 Err(why) => panic!("couldn't create {}: {}", output_path.display(), why),
             };
             let mut bw = BufWriter::new(output_file);
-            for (var, val) in &witness_map {
-                write!(&mut bw, "{} {}\n", var, val.to_dec_string())
-                    .expect("Unable to write data to file.");
-            }
+            write!(&mut bw, "{}", witness).expect("Unable to write data to file.");
             bw.flush().expect("Unable to flush buffer.");
         }
         #[cfg(feature = "libsnark")]
@@ -656,7 +643,7 @@ mod tests {
 
             let (..) = r1cs_program(program_flattened.clone());
             let _ = program_flattened
-                .execute(vec![FieldPrime::from(0)])
+                .execute(&vec![FieldPrime::from(0)])
                 .unwrap();
         }
     }
@@ -690,7 +677,7 @@ mod tests {
 
             let result = std::panic::catch_unwind(|| {
                 let _ = program_flattened
-                    .execute(vec![FieldPrime::from(0)])
+                    .execute(&vec![FieldPrime::from(0)])
                     .unwrap();
             });
             assert!(result.is_err());
diff --git a/zokrates_core/src/ir/interpreter.rs b/zokrates_core/src/ir/interpreter.rs
index 99c0f58a..f7b346b8 100644
--- a/zokrates_core/src/ir/interpreter.rs
+++ b/zokrates_core/src/ir/interpreter.rs
@@ -3,17 +3,57 @@ use ir::*;
 use std::collections::BTreeMap;
 use zokrates_field::field::Field;
 
+pub type ExecutionResult<T> = Result<Witness<T>, Error>;
+
+pub struct Witness<T: Field>(BTreeMap<FlatVariable, T>);
+
+impl<T: Field> Witness<T> {
+    pub fn return_values(&self) -> Vec<T> {
+        self.0
+            .clone()
+            .into_iter()
+            .filter(|(k, _)| k.is_output())
+            .map(|(_, v)| v)
+            .collect()
+    }
+
+    pub fn format_outputs(&self) -> String {
+        self.0
+            .iter()
+            .filter_map(|(variable, value)| match variable {
+                variable if variable.is_output() => Some(format!("{} {}", variable, value)),
+                _ => None,
+            })
+            .collect::<Vec<String>>()
+            .join("\n")
+    }
+}
+
+impl<T: Field> fmt::Display for Witness<T> {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(
+            f,
+            "{}",
+            self.0
+                .iter()
+                .map(|(k, v)| format!("{} {}", k, v.to_dec_string()))
+                .collect::<Vec<_>>()
+                .join("\n")
+        )
+    }
+}
+
 impl<T: Field> Prog<T> {
-    pub fn execute(self, inputs: Vec<T>) -> Result<BTreeMap<FlatVariable, T>, Error<T>> {
-        let main = self.main;
-        assert_eq!(main.arguments.len(), inputs.len());
+    pub fn execute<U: Into<T> + Clone>(&self, inputs: &Vec<U>) -> ExecutionResult<T> {
+        let main = &self.main;
+        self.check_inputs(&inputs)?;
         let mut witness = BTreeMap::new();
         witness.insert(FlatVariable::one(), T::one());
         for (arg, value) in main.arguments.iter().zip(inputs.iter()) {
-            witness.insert(arg.clone(), value.clone());
+            witness.insert(arg.clone(), value.clone().into());
         }
 
-        for statement in main.statements {
+        for statement in &main.statements {
             match statement {
                 Statement::Constraint(quad, lin) => match lin.is_assignee(&witness) {
                     true => {
@@ -24,7 +64,10 @@ impl<T: Field> Prog<T> {
                         let lhs_value = quad.evaluate(&witness);
                         let rhs_value = lin.evaluate(&witness);
                         if lhs_value != rhs_value {
-                            return Err(Error::Constraint(quad, lin, lhs_value, rhs_value));
+                            return Err(Error::UnsatisfiedConstraint {
+                                left: lhs_value.to_dec_string(),
+                                right: rhs_value.to_dec_string(),
+                            });
                         }
                     }
                 },
@@ -44,7 +87,18 @@ impl<T: Field> Prog<T> {
             }
         }
 
-        Ok(witness)
+        Ok(Witness(witness))
+    }
+
+    fn check_inputs<U>(&self, inputs: &Vec<U>) -> Result<(), Error> {
+        if self.main.arguments.len() == inputs.len() {
+            Ok(())
+        } else {
+            Err(Error::WrongInputCount {
+                expected: self.main.arguments.len(),
+                received: inputs.len(),
+            })
+        }
     }
 }
 
@@ -69,21 +123,35 @@ impl<T: Field> QuadComb<T> {
     }
 }
 
-#[derive(PartialEq, Debug)]
-pub enum Error<T: Field> {
-    Constraint(QuadComb<T>, LinComb<T>, T, T),
+#[derive(PartialEq, Serialize, Deserialize)]
+pub enum Error {
+    UnsatisfiedConstraint { left: String, right: String },
     Solver,
+    WrongInputCount { expected: usize, received: usize },
 }
 
-impl<T: Field> fmt::Display for Error<T> {
+impl fmt::Display for Error {
     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
         match *self {
-            Error::Constraint(ref quad, ref lin, ref left_value, ref right_value) => write!(
-                f,
-                "Expected {} to equal {}, but {} != {}",
-                quad, lin, left_value, right_value
-            ),
+            Error::UnsatisfiedConstraint {
+                ref left,
+                ref right,
+            } => write!(f, "Expected {} to equal {}", left, right),
             Error::Solver => write!(f, ""),
+            Error::WrongInputCount { expected, received } => write!(
+                f,
+                "Program takes {} input{} but was passed {} value{}",
+                expected,
+                if expected == 1 { "" } else { "s" },
+                received,
+                if received == 1 { "" } else { "s" }
+            ),
         }
     }
 }
+
+impl fmt::Debug for Error {
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        write!(f, "{}", self)
+    }
+}
diff --git a/zokrates_core/src/ir/mod.rs b/zokrates_core/src/ir/mod.rs
index 3d81bed9..62b45cb1 100644
--- a/zokrates_core/src/ir/mod.rs
+++ b/zokrates_core/src/ir/mod.rs
@@ -13,6 +13,9 @@ mod interpreter;
 use self::expression::LinComb;
 use self::expression::QuadComb;
 
+pub use self::interpreter::Error;
+pub use self::interpreter::ExecutionResult;
+
 #[derive(Debug, Serialize, Deserialize, Clone)]
 pub enum Statement<T: Field> {
     Constraint(QuadComb<T>, LinComb<T>),
diff --git a/zokrates_core/tests/bench/add.code b/zokrates_core/tests/bench/add.code
new file mode 100644
index 00000000..d2277712
--- /dev/null
+++ b/zokrates_core/tests/bench/add.code
@@ -0,0 +1,2 @@
+def main(field a, field b) -> (field):
+	return a + b
\ No newline at end of file
diff --git a/zokrates_core/tests/bench/add.json b/zokrates_core/tests/bench/add.json
new file mode 100644
index 00000000..f7407d67
--- /dev/null
+++ b/zokrates_core/tests/bench/add.json
@@ -0,0 +1,27 @@
+{
+	"tests": [
+		{
+			"input": {
+				"values": ["1", "2"]
+			},
+			"output": {
+				"Ok": {
+					"values": ["3"]
+				}
+			}
+		},
+		{
+			"input": {
+				"values": ["1", "2", "42"]
+			},
+			"output": {
+				"Err": {
+					"WrongInputCount": {
+						"expected": 2,
+						"received": 3
+					}
+				}
+			}
+		}
+	]
+}
\ No newline at end of file
diff --git a/zokrates_core/tests/bench/assert_one.code b/zokrates_core/tests/bench/assert_one.code
new file mode 100644
index 00000000..03dafa87
--- /dev/null
+++ b/zokrates_core/tests/bench/assert_one.code
@@ -0,0 +1,3 @@
+def main(field a) -> (field):
+	a == 1
+	return 1
\ No newline at end of file
diff --git a/zokrates_core/tests/bench/assert_one.json b/zokrates_core/tests/bench/assert_one.json
new file mode 100644
index 00000000..5709a87f
--- /dev/null
+++ b/zokrates_core/tests/bench/assert_one.json
@@ -0,0 +1,17 @@
+{
+	"tests": [
+		{
+			"input": {
+				"values": ["0"]
+			},
+			"output": {
+				"Err": {
+					"UnsatisfiedConstraint": {
+						"left": "1",
+						"right": "0"
+					}
+				}
+			}
+		}
+	]
+}
\ No newline at end of file
diff --git a/zokrates_core/tests/integration.rs b/zokrates_core/tests/integration.rs
new file mode 100644
index 00000000..af824ffc
--- /dev/null
+++ b/zokrates_core/tests/integration.rs
@@ -0,0 +1,12 @@
+extern crate serde_json;
+extern crate zokrates_core;
+#[macro_use]
+extern crate serde_derive;
+
+#[macro_use]
+mod utils;
+
+zokrates_test! {
+    add,
+    assert_one,
+}
diff --git a/zokrates_core/tests/utils/mod.rs b/zokrates_core/tests/utils/mod.rs
new file mode 100644
index 00000000..b66ce26d
--- /dev/null
+++ b/zokrates_core/tests/utils/mod.rs
@@ -0,0 +1,120 @@
+extern crate serde_json;
+
+use std::io;
+use zokrates_core::compile::{compile as generic_compile, CompileError};
+use zokrates_core::field::{Field, FieldPrime};
+use zokrates_core::ir;
+
+#[derive(Serialize, Deserialize)]
+pub struct Tests {
+    pub tests: Vec<Test>,
+}
+
+#[derive(Serialize, Deserialize)]
+pub struct Input {
+    pub values: Vec<Val>,
+}
+
+#[derive(Serialize, Deserialize)]
+pub struct Test {
+    pub input: Input,
+    pub output: TestResult,
+}
+
+pub type TestResult = Result<Output, ir::Error>;
+
+#[derive(PartialEq, Debug)]
+pub struct ComparableResult(Result<Vec<FieldPrime>, ir::Error>);
+
+#[derive(Serialize, Deserialize)]
+pub struct Output {
+    values: Vec<Val>,
+}
+
+type Val = String;
+
+impl From<ir::ExecutionResult<FieldPrime>> for ComparableResult {
+    fn from(r: ir::ExecutionResult<FieldPrime>) -> ComparableResult {
+        ComparableResult(r.map(|v| v.return_values()))
+    }
+}
+
+impl From<TestResult> for ComparableResult {
+    fn from(r: TestResult) -> ComparableResult {
+        ComparableResult(r.map(|v| {
+            v.values
+                .into_iter()
+                .map(|v| FieldPrime::from_dec_string(v))
+                .collect()
+        }))
+    }
+}
+
+pub fn compare(
+    result: ir::ExecutionResult<FieldPrime>,
+    expected: TestResult,
+) -> Result<(), String> {
+    // extract outputs from result
+    let result = ComparableResult::from(result);
+    // deserialize expected result
+    let expected = ComparableResult::from(expected);
+
+    if result != expected {
+        return Err(format!(
+            "Expected {:?} but found {:?}",
+            expected.0, result.0
+        ));
+    }
+
+    Ok(())
+}
+
+pub fn read_file(path: &str) -> String {
+    use std::fs::File;
+    use std::io::Read;
+    let mut file = File::open(format!("./tests/bench/{}", path)).expect("Unable to open the file");
+    let mut contents = String::new();
+    file.read_to_string(&mut contents)
+        .expect("Unable to read the file");
+
+    contents
+}
+
+pub fn compile(code: &str) -> Result<ir::Prog<FieldPrime>, CompileError<FieldPrime>> {
+    generic_compile::<FieldPrime, &[u8], &[u8], io::Error>(&mut code.as_bytes(), None, None)
+}
+
+macro_rules! zokrates_test {
+    ($($name:ident,)*) => {
+          $(
+            #[test]
+            fn $name() {
+
+                use zokrates_core::field::{FieldPrime, Field};
+
+                let code_string = $crate::utils::read_file(&format!("./{}.code", stringify!($name)));
+                let test_string = $crate::utils::read_file(&format!("./{}.json", stringify!($name)));
+
+                let bin = $crate::utils::compile(&code_string).unwrap();
+
+                let t: $crate::utils::Tests = serde_json::from_str(&test_string).unwrap();
+
+                for test in t.tests.into_iter() {
+                    let input = &test.input.values;
+                    let output = bin.execute(&input.iter().map(|v| FieldPrime::from_dec_string(v.clone())).collect());
+
+                    let context = format!("
+{}
+
+Called with input ({})
+            ", code_string, input.iter().map(|i| format!("{}", i)).collect::<Vec<_>>().join(", "));
+
+                    match $crate::utils::compare(output, test.output) {
+                        Err(e) => panic!("{}{}", context, e),
+                        Ok(..) => {}
+                    };
+                }
+            }
+          )*
+    };
+}