snarkjs

This is a JavaScript and Pure Web Assembly implementation of zkSNARK and PLONK schemes. It uses the Groth16 Protocol (3 points only and 3 pairings), PLONK, and FFLONK.

This library includes all the tools required to perform trusted setup multi-party ceremonies: including the universal powers of tau ceremony, and the second phase circuit-specific ceremonies.

Any zk-snark project can pick a round from the common phase 1 to start their circuit-specific phase 2 ceremony.

The formats used in this library for the multi-party computation are compatible with the ones used in Semaphore’s Perpetual Powers of Tau and other implementations.

This library uses the compiled circuits generated by the circom compiler.

It works in node.js as well as directly in the browser.

It’s an ES module, so it can be directly imported into bigger projects using Rollup or Webpack.

The low-level cryptography is performed directly in wasm, and uses worker threads to parallelize the computations. The result is a high-performance library with benchmarks comparable to host implementations.

Preliminaries

Install node

First off, make sure you have a recent LTS version of Node.js installed. Non-LTS and versions prior to v18 are not guaranteed to work.

If you’re not sure which version of Node you have installed, you can run:

node -v

To download the latest version of Node, see here.

Install snarkjs

To install snarkjs run:

npm install -g snarkjs@latest

If you’re seeing an error, try prefixing both commands with sudo and running them again.

Understand the help command

To see a list of all snarkjs commands, as well as descriptions of their inputs and outputs, run:

snarkjs --help

You can also use the --help option with specific commands:

snarkjs groth16 prove --help

Most of the commands have an alternative shorter alias (which you can discover using --help).

For example, the previous command can also be invoked with:

snarkjs g16p --help

Debugging tip

If you feel a command is taking longer than it should, re-run it with a -v or --verbose option to see more details about how it’s progressing and where it’s getting blocked.

Install circom

To install circom, follow the instructions at installing circom.

Guide

0. Create and move into a new directory

mkdir snarkjs_example
cd snarkjs_example

1. Start a new powers of tau ceremony

snarkjs powersoftau new bn128 14 pot14_0000.ptau -v

The new command is used to start a powers of tau ceremony.

The first parameter after new refers to the type of curve you wish to use. At the moment, we support both bn128 and bls12-381.

The second parameter, in this case 14, is the power of two of the maximum number of constraints that the ceremony can accept: in this case, the number of constraints is 2 ^ 14 = 16,384. The maximum value supported here is 28, which means you can use snarkjs to securely generate zk-snark parameters for circuits with up to 2 ^ 28 (≈268 million) constraints.

2. Contribute to the ceremony

snarkjs powersoftau contribute pot14_0000.ptau pot14_0001.ptau --name="First contribution" -v

The contribute command creates a ptau file with a new contribution.

You’ll be prompted to enter some random text to provide an extra source of entropy.

contribute takes as input the transcript of the protocol so far, in this case pot14_0000.ptau, and outputs a new transcript, in this case pot14_0001.ptau, which includes the computation carried out by the new contributor (ptau files contain a history of all the challenges and responses that have taken place so far).

name can be anything you want, and is just included for reference (it will be printed when you verify the file (step 5).

3. Provide a second contribution

snarkjs powersoftau contribute pot14_0001.ptau pot14_0002.ptau --name="Second contribution" -v -e="some random text"

By letting you write the random text as part of the command, the -e parameter allows contribute to be non-interactive.

4. Provide a third contribution using third-party software

snarkjs powersoftau export challenge pot14_0002.ptau challenge_0003
snarkjs powersoftau challenge contribute bn128 challenge_0003 response_0003 -e="some random text"
snarkjs powersoftau import response pot14_0002.ptau response_0003 pot14_0003.ptau -n="Third contribution name"

The challenge and response files are compatible with this software.

This allows you to use different types of software in a single ceremony.

5. Verify the protocol so far

snarkjs powersoftau verify pot14_0003.ptau

The verify command verifies a ptau (powers of tau) file, which means it checks all the contributions to the multi-party computation (MPC) up to that point. It also prints the hashes of all the intermediate results to the console.

If everything checks out, you should see the following at the top of the output:

[INFO]  snarkJS: Powers Of tau file OK!

In sum, whenever a new zk-snark project needs to perform a trusted setup, you can just pick the latest ptau file, and run the verify command to verify the entire chain of challenges and responses so far.

6. Apply a random beacon

snarkjs powersoftau beacon pot14_0003.ptau pot14_beacon.ptau 0102030405060708090a0b0c0d0e0f101112131415161718191a1b1c1d1e1f 10 -n="Final Beacon"

The beacon command creates a ptau file with a contribution applied in the form of a random beacon.

We need to apply a random beacon in order to finalize phase 1 of the trusted setup.

To paraphrase Sean Bowe and Ariel Gabizon, a random beacon is a source of public randomness that is not available before a fixed time. The beacon itself can be a delayed hash function (e.g. 2^40 iterations of SHA256) evaluated on some high entropy and publicly available data. Possible sources of data include: the closing value of the stock market on a certain date in the future, the output of a selected set of national lotteries, or the value of a block at a particular height in one or more blockchains. E.g. the hash of the 11 millionth Ethereum block (which as of this writing is some 3 months in the future). See here for more on the importance of a random beacon.

For the purposes of this tutorial, the beacon is essentially a delayed hash function evaluated on 0102030405060708090a0b0c0d0e0f101112131415161718191a1b1c1d1e1f (in practice this value will be some form of high entropy and publicly available data of your choice). The next input – in our case 10 – just tells snarkjs to perform 2 ^ 10 iterations of this hash function.

Note that security holds even if an adversary has limited influence on the beacon.

7. Prepare phase 2

snarkjs powersoftau prepare phase2 pot14_beacon.ptau pot14_final.ptau -v

We’re now ready to prepare phase 2 of the setup (the circuit-specific phase).

Under the hood, the prepare phase2 command calculates the encrypted evaluation of the Lagrange polynomials at tau for tau, alpha*tau, and beta*tau. It takes the beacon ptau file we generated in the previous step and outputs a final ptau file which will be used to generate the circuit proving and verification keys.

NOTE

Prepared (phase2) Ptau files for bn128 with 54 contributions and a beacon can be found here:

power maxConstraints file hash 8 256 powersOfTau28_hez_final_08.ptau d6a8fb3a04feb600096c3b791f936a578c4e664d262e4aa24beed1b7a9a96aa5eb72864d628db247e9293384b74b36ffb52ca8d148d6e1b8b51e279fdf57b583 9 512 powersOfTau28_hez_final_09.ptau 94f108a80e81b5d932d8e8c9e8fd7f46cf32457e31462deeeef37af1b71c2c1b3c71fb0d9b59c654ec266b042735f50311f9fd1d4cadce47ab234ad163157cb5 10 1k powersOfTau28_hez_final_10.ptau 6cfeb8cda92453099d20120bdd0e8a5c4e7706c2da9a8f09ccc157ed2464d921fd0437fb70db42104769efd7d6f3c1f964bcf448c455eab6f6c7d863e88a5849 11 2k powersOfTau28_hez_final_11.ptau 47c282116b892e5ac92ca238578006e31a47e7c7e70f0baa8b687f0a5203e28ea07bbbec765a98dcd654bad618475d4661bfaec3bd9ad2ed12e7abc251d94d33 12 4k powersOfTau28_hez_final_12.ptau ded2694169b7b08e898f736d5de95af87c3f1a64594013351b1a796dbee393bd825f88f9468c84505ddd11eb0b1465ac9b43b9064aa8ec97f2b73e04758b8a4a 13 8k powersOfTau28_hez_final_13.ptau 58efc8bf2834d04768a3d7ffcd8e1e23d461561729beaac4e3e7a47829a1c9066d5320241e124a1a8e8aa6c75be0ba66f65bc8239a0542ed38e11276f6fdb4d9 14 16k powersOfTau28_hez_final_14.ptau eeefbcf7c3803b523c94112023c7ff89558f9b8e0cf5d6cdcba3ade60f168af4a181c9c21774b94fbae6c90411995f7d854d02ebd93fb66043dbb06f17a831c1 15 32k powersOfTau28_hez_final_15.ptau 982372c867d229c236091f767e703253249a9b432c1710b4f326306bfa2428a17b06240359606cfe4d580b10a5a1f63fbed499527069c18ae17060472969ae6e 16 64k powersOfTau28_hez_final_16.ptau 6a6277a2f74e1073601b4f9fed6e1e55226917efb0f0db8a07d98ab01df1ccf43eb0e8c3159432acd4960e2f29fe84a4198501fa54c8dad9e43297453efec125 17 128k powersOfTau28_hez_final_17.ptau 6247a3433948b35fbfae414fa5a9355bfb45f56efa7ab4929e669264a0258976741dfbe3288bfb49828e5df02c2e633df38d2245e30162ae7e3bcca5b8b49345 18 256k powersOfTau28_hez_final_18.ptau 7e6a9c2e5f05179ddfc923f38f917c9e6831d16922a902b0b4758b8e79c2ab8a81bb5f29952e16ee6c5067ed044d7857b5de120a90704c1d3b637fd94b95b13e 19 512k powersOfTau28_hez_final_19.ptau bca9d8b04242f175189872c42ceaa21e2951e0f0f272a0cc54fc37193ff6648600eaf1c555c70cdedfaf9fb74927de7aa1d33dc1e2a7f1a50619484989da0887 20 1M powersOfTau28_hez_final_20.ptau 89a66eb5590a1c94e3f1ee0e72acf49b1669e050bb5f93c73b066b564dca4e0c7556a52b323178269d64af325d8fdddb33da3a27c34409b821de82aa2bf1a27b 21 2M powersOfTau28_hez_final_21.ptau 9aef0573cef4ded9c4a75f148709056bf989f80dad96876aadeb6f1c6d062391f07a394a9e756d16f7eb233198d5b69407cca44594c763ab4a5b67ae73254678 22 4M powersOfTau28_hez_final_22.ptau 0d64f63dba1a6f11139df765cb690da69d9b2f469a1ddd0de5e4aa628abb28f787f04c6a5fb84a235ec5ea7f41d0548746653ecab0559add658a83502d1cb21b 23 8M powersOfTau28_hez_final_23.ptau 3063a0bd81d68711197c8820a92466d51aeac93e915f5136d74f63c394ee6d88c5e8016231ea6580bec02e25d491f319d92e77f5c7f46a9caa8f3b53c0ea544f 24 16M powersOfTau28_hez_final_24.ptau fa404d140d5819d39984833ca5ec3632cd4995f81e82db402371a4de7c2eae8687c62bc632a95b0c6aadba3fb02680a94e09174b7233ccd26d78baca2647c733 25 32M powersOfTau28_hez_final_25.ptau 0377d860cdb09a8a31ea1b0b8c04335614c8206357181573bf294c25d5ca7dff72387224fbd868897e6769f7805b3dab02854aec6d69d7492883b5e4e5f35eeb 26 64M powersOfTau28_hez_final_26.ptau 418dee4a74b9592198bd8fd02ad1aea76f9cf3085f206dfd7d594c9e264ae919611b1459a1cc920c2f143417744ba9edd7b8d51e44be9452344a225ff7eead19 27 128M powersOfTau28_hez_final_27.ptau 10ffd99837c512ef99752436a54b9810d1ac8878d368fb4b806267bdd664b4abf276c9cd3c4b9039a1fa4315a0c326c0e8e9e8fe0eb588ffd4f9021bf7eae1a1 28 256M powersOfTau28_hez_final.ptau 55c77ce8562366c91e7cda394cf7b7c15a06c12d8c905e8b36ba9cf5e13eb37d1a429c589e8eaba4c591bc4b88a0e2828745a53e170eac300236f5c1a326f41a

There is a file truncated for each power of two.

The complete file is powersOfTau28_hez_final.ptau which includes 2**28 powers.

And it’s blake2b hash is:

55c77ce8562366c91e7cda394cf7b7c15a06c12d8c905e8b36ba9cf5e13eb37d1a429c589e8eaba4c591bc4b88a0e2828745a53e170eac300236f5c1a326f41a

You can find more information about the ceremony here

The last ptau file was generated using this procedure:

https://www.reddit.com/r/ethereum/comments/iftos6/powers_of_tau_selection_for_hermez_rollup/

8. Verify the final ptau

snarkjs powersoftau verify pot14_final.ptau

The verify command verifies a powers of tau file.

Before we go ahead and create the circuit, we perform a final check and verify the final protocol transcript.

Notice there is no longer a warning informing you that the file does not contain phase 2 precalculated values.

9. Create the circuit

cat <<EOT > circuit.circom
pragma circom 2.0.0;

template Multiplier(n) {
    signal input a;
    signal input b;
    signal output c;

    signal int[n];

    int[0] <== a*a + b;
    for (var i=1; i<n; i++) {
    int[i] <== int[i-1]*int[i-1] + b;
    }

    c <== int[n-1];
}

component main = Multiplier(1000);
EOT

We create a circom file that allows us to easily test the system with a different number of constraints.

In this case, we’ve chosen 1000, but we can change this to anything we want (as long as the value we choose is below the number we defined in step 1).

10. Compile the circuit

circom --r1cs --wasm --c --sym --inspect circuit.circom

The circom command takes one input (the circuit to compile, in our case circuit.circom) and three options:

• r1cs: generates circuit.r1cs (the r1cs constraint system of the circuit in binary format).

• wasm: generates circuit.wasm (the wasm code to generate the witness – more on that later).

• c: generates c++ witness calculator code.

• sym: generates circuit.sym (a symbols file required for debugging and printing the constraint system in an annotated mode).

• inspect: does additional checks over the constraints produced.

11. View information about the circuit

snarkjs r1cs info circuit.r1cs

The info command is used to print circuit stats.

You should see the following output:

[INFO]  snarkJS: Curve: bn-128
[INFO]  snarkJS: # of Wires: 1003
[INFO]  snarkJS: # of Constraints: 1000
[INFO]  snarkJS: # of Private Inputs: 2
[INFO]  snarkJS: # of Public Inputs: 0
[INFO]  snarkJS: # of Outputs: 1

This information fits with our mental map of the circuit we created: we had two private inputs a and b, one output c, and a thousand constraints of the form a * b = c.

12. Print the constraints

snarkjs r1cs print circuit.r1cs circuit.sym

To double-check, we print the constraints of the circuit.

You should see a thousand constraints of the form:

[ -main.int[i] ] * [ main.int[i] ] - [ main.b -main.int[i+1] ] = 0

13. Export r1cs to json

snarkjs r1cs export json circuit.r1cs circuit.r1cs.json
cat circuit.r1cs.json

We export r1cs to json format to make it human-readable.

14. Calculate the witness

First, we create a file with the inputs for our circuit:

cat <<EOT > input.json
{"a": "3", "b": "11"}
EOT

Note that integers in json file are enclosed in double quotation marks, because otherwise json format loses precision when working with big integers.

Now, we use the Javascript/WASM program created by circom in the directory _js to create the witness (values of all the wires) for our inputs:

snarkjs wtns calculate circuit_js/circuit.wasm input.json witness.wtns

Alternatively, you can do:

node circuit_js/generate_witness.js circuit_js/circuit.wasm input.json witness.wtns

We can check if the generated witness complies with the r1cs file with the following command:

snarkjs wtns check circuit.r1cs witness.wtns

15. Setup

Currently, snarkjs supports 3 proving systems: Groth16, PLONK, and FFLONK (Beta version).

Groth16 requires a trusted ceremony for each circuit. PLONK and FFLONK do not require it, it’s enough with the powers of tau ceremony which is universal.

Plonk

snarkjs plonk setup circuit.r1cs pot14_final.ptau circuit_final.zkey

Fflonk

snarkjs fflonk setup circuit.r1cs pot14_final.ptau circuit.zkey

You can jump directly to Section 21 as PLONK and FFLONK do not require a circuit-specific trusted ceremony.

Groth16

snarkjs groth16 setup circuit.r1cs pot14_final.ptau circuit_0000.zkey

This generates the reference zkey without phase 2 contributions.

IMPORTANT: Do not use this zkey in production, as it’s not safe. It requires at least one contribution.

The zkey new command creates an initial zkey file with zero contributions.

The zkey is a zero-knowledge key that includes both the proving and verification keys as well as phase 2 contributions.

Importantly, one can verify whether a zkey belongs to a specific circuit or not.

Note that circuit_0000.zkey (the output of the zkey command above) does not include any contributions yet, so it cannot be used in a final circuit.

The following steps (15-20) are similar to the equivalent phase 1 steps, except we use zkey instead of powersoftau as the main command, and we generate zkey rather than ptau files.

16. Contribute to the phase 2 ceremony

snarkjs zkey contribute circuit_0000.zkey circuit_0001.zkey --name="1st Contributor Name" -v

The zkey contribute command creates a zkey file with a new contribution.

As in phase 1, you’ll be prompted to enter some random text to provide an extra source of entropy.

17. Provide a second contribution

snarkjs zkey contribute circuit_0001.zkey circuit_0002.zkey --name="Second contribution Name" -v -e="Another random entropy"

We provide a second contribution.

18. Provide a third contribution using third-party software

snarkjs zkey export bellman circuit_0002.zkey  challenge_phase2_0003
snarkjs zkey bellman contribute bn128 challenge_phase2_0003 response_phase2_0003 -e="some random text"
snarkjs zkey import bellman circuit_0002.zkey response_phase2_0003 circuit_0003.zkey -n="Third contribution name"

And a third using third-party software.

19. Verify the latest zkey

snarkjs zkey verify circuit.r1cs pot14_final.ptau circuit_0003.zkey

The zkey verify command verifies a zkey file. It also prints the hashes of all the intermediary results to the console.

We verify the zkey file we created in the previous step, which means we check all the contributions to the second phase of the multi-party computation (MPC) up to that point.

This command also checks that the zkey file matches the circuit.

If everything checks out, you should see the following:

[INFO]  snarkJS: ZKey Ok!

20. Apply a random beacon

snarkjs zkey beacon circuit_0003.zkey circuit_final.zkey 0102030405060708090a0b0c0d0e0f101112131415161718191a1b1c1d1e1f 10 -n="Final Beacon phase2"

The zkey beacon command creates a zkey file with a contribution applied in the form of a random beacon.

We use it to apply a random beacon to the latest zkey after the final contribution has been made (this is necessary to generate a final zkey file and finalize phase 2 of the trusted setup).

21. Verify the final zkey

snarkjs zkey verify circuit.r1cs pot14_final.ptau circuit_final.zkey

Before we go ahead and export the verification key as a json, we perform a final check and verify the final protocol transcript (zkey).

22. Export the verification key

snarkjs zkey export verificationkey circuit_final.zkey verification_key.json

We export the verification key from circuit_final.zkey into verification_key.json.

23. Create the proof

PLONK

snarkjs plonk prove circuit_final.zkey witness.wtns proof.json public.json

FFLONK

snarkjs fflonk prove circuit.zkey witness.wtns proof.json public.json

Groth16

snarkjs groth16 prove circuit_final.zkey witness.wtns proof.json public.json

We create the proof. This command generates the files proof.json and public.json:

• proof.json contains the actual proof.
• public.json contains the values of the public inputs and output.

23a. Calculate the witness and generate the proof in one step

Note that it’s also possible to create the proof and calculate the witness in the same command by running:

PLONK

snarkjs plonk fullprove witness.json circuit.wasm circuit_final.zkey proof.json public.json

FFLONK

snarkjs fflonk fullprove witness.json circuit.wasm circuit_final.zkey proof.json public.json

Groth16

snarkjs groth16 fullprove input.json circuit.wasm circuit_final.zkey proof.json public.json

24. Verify the proof

PLONK

snarkjs plonk verify verification_key.json public.json proof.json

FFLONK

snarkjs fflonk verify verification_key.json public.json proof.json

Groth16

snarkjs groth16 verify verification_key.json public.json proof.json

We use this command to verify the proof, passing in the verification_key we exported earlier.

If all is well, you should see that OK has been outputted to your console. This signifies the proof is valid.

25. Turn the verifier into a smart contract

snarkjs zkey export solidityverifier circuit_final.zkey verifier.sol

Finally, we export the verifier as a Solidity smart contract so that we can publish it on-chain – using remix for example. For details on how to do this, refer to section 4 of this tutorial.

26. Simulate a verification call

snarkjs zkey export soliditycalldata public.json proof.json

We use soliditycalldata to simulate a verification call and cut-and-paste the result directly into the verifyProof field in the deployed smart contract in the remix environment.

And voila! That’s all there is to it :)

Using Node

npm init
npm install snarkjs

const snarkjs = require("snarkjs");
const fs = require("fs");

async function run() {
    const { proof, publicSignals } = await snarkjs.groth16.fullProve({a: 10, b: 21}, "circuit.wasm", "circuit_final.zkey");

    console.log("Proof: ");
    console.log(JSON.stringify(proof, null, 1));

    const vKey = JSON.parse(fs.readFileSync("verification_key.json"));

    const res = await snarkjs.groth16.verify(vKey, publicSignals, proof);

    if (res === true) {
        console.log("Verification OK");
    } else {
        console.log("Invalid proof");
    }

}

run().then(() => {
    process.exit(0);
});

In the browser

Load snarkjs.min.js and start using it as usual.

cp node_modules/snarkjs/build/snarkjs.min.js .

<!doctype html>
<html>
<head>
  <title>Snarkjs client example</title>
</head>
<body>

  <h1>Snarkjs client example</h1>
  <button id="bGenProof"> Create proof </button>

  <!-- JS-generated output will be added here. -->
  <pre class="proof"> Proof: <code id="proof"></code></pre>

  <pre class="proof"> Result: <code id="result"></code></pre>


  <script src="snarkjs.min.js">   </script>


  <!-- This is the bundle generated by rollup.js -->
  <script>

const proofComponent = document.getElementById('proof');
const resultComponent = document.getElementById('result');
const bGenProof = document.getElementById("bGenProof");

bGenProof.addEventListener("click", calculateProof);

async function calculateProof() {

    const { proof, publicSignals } =
      await snarkjs.groth16.fullProve( { a: 3, b: 11}, "circuit.wasm", "circuit_final.zkey");

    proofComponent.innerHTML = JSON.stringify(proof, null, 1);


    const vkey = await fetch("verification_key.json").then( function(res) {
        return res.json();
    });

    const res = await snarkjs.groth16.verify(vkey, publicSignals, proof);

    resultComponent.innerHTML = res;
}

  </script>

</body>
</html>

Security

We advise using the latest version of snarkjs as previous versions may contain bugs and security issues.

Common Issues

Multithreading with Worker Threads is not available in my execution environment

This happens in Bun, browser extensions, SES envs, etc.

You can switch to singlethreaded proof calculation using the following code:

const result = await snarkjs.groth16.prove(zkey_final, wtns, undefined, {singleThread: true});

Control initial memory allocation in witness calculator

To minimize amount of memory allocated on the web you could pass a following option:

await wtnsCalculate(input, wasmFile, wtns, {memorySize: 0});

Issues with Web Pack

Further resources

• Announcing the Perpetual Powers of Tau Ceremony to benefit all zk-SNARK projects
• Scalable Multi-party Computation for zk-SNARK Parameters in the Random Beacon Model
• phase2-bn254
• Perpetual Powers of Tau
• Powers of Tau
• Trusted setup ceremonies explored
• Simple react project using snarkjs

Final note

We hope you enjoyed this quick walk-through. Please address any questions you may have to our telegram group (it’s also a great way to join the community and stay up-to-date with the latest circom and snarkjs developments) 💙

License

snarkjs is part of the iden3 project copyright 2018 0KIMS association and published with GPL-3 license. Please check the COPYING file for more details.