Command line interface for borrowing and lending crypto-currency using Ethereum smart contracts
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
bin integrated dharma-credit and pointing to kovan now by default Aug 8, 2017
examples fixed exponentiation error Nov 9, 2017
lib fixed exponentiation error Nov 9, 2017
node_modules added transform generate runtime Aug 25, 2017
src fixed exponentiation error Nov 9, 2017
.babelrc added transform generate runtime Aug 25, 2017
.gitignore cleaned up dependencies Jul 25, 2017
.mockyeah init Jul 8, 2017 Update Jun 1, 2018
package-lock.json 0.0.29 Aug 25, 2017
package.json 0.0.29 Aug 25, 2017
yarn.lock bumped dharma-client to 0.0.18 Aug 8, 2017

NOTE: This repository relies on a deprecated set of smart contracts used in Dharma protocol -- the most current set of smart contracts in the Dharma protocol are at This repository is kept public for historical purposes, but we don't recommend using the tool for any serious production use case.

Dharma Command Line Interface

Dharma is a protocol for unsecured, peer-to-peer cryptocurrency lending built on top of Ethereum smart contracts. Our first testnet release is the Dharma CLI -- a command line tool that makes it easy to:

  1. Access a programmatic line of crypto credit from the command line in under 5 minutes

  1. Build bots that auto-invest in a portfolio of loans under hyper-customizable, programmable criteria

See installation instructions below to give it a try, and check out the Dharma Protocol White Paper for more extensive information on the specifics of the protocol.

The Dharma CLI runs on a private Ethereum Testnet spun up specifically for demonstration purposes. However, nothing about the protocol or implementation requires usage of permissioned or private chains, and we intend on moving the implementation to a public chain in the near future.

The Dharma CLI is live on Kovan Testnet as of release 0.0.25! Older versions of the CLI utilized a private Ethereum Testnet spun up specifically for demonstration purposes -- for any troubleshooting with the deprecated CLI, please email

This is highly experimental, alpha-stage ~mAd~SciEncE~, so any and all bug reports, uncompromising feedback, and issue reports are welcome.

Table of Contents

  1. Installation
  2. Usage
  3. Overview
  4. How It Works
  5. Decision Engine Documentation


The Dharma CLI runs in a node environment. In order to install the CLI, run the following in your command line:

# The following two commands are necessary in order to install scrypt for the first time
$ npm install -g node-gyp
$ npm install scrypt

$ npm install -g dharma-cli

If you run into installation errors related to the scrypt package, make sure you have g++ installed and configured locally. If you run into any other installation snags, please open up an issue and I'd be happy to take a look and assist you.


For borrowers...

$ dharma borrow

The CLI will guide you through a wizard that in which you will generate a client-side wallet, verify your identity, and then decide on the loan parameters you desire.

For lenders...

$ dharma invest <decisionEnginePath>

The CLI will start up a daemon that automatically bids on loans in the Dharma network according to the criteria stipulated by the Javascript file pointed to by decisionEnginePath. The CLI will open a dashboard that allows investors to get a live snapshot of their portfolio over time.

For example, one could write a decision engine that bids on loans on the basis of a simple, fixed risk-to-interest ratio:


class DecisionEngine {
  constructor(web3) {
    this.web3 = web3;

  async decide(loan) {
    if (
      return false;

    return {
      amount: loan.principal,
      minInterestRate: loan.defaultRisk.times(RISK_TO_INTEREST_RATIO)

Or, alternatively, one could write a decision engine that bids on loans on the basis of current macro interest rates, benchmark data from existing online lenders, the Pisces horoscope on any given day -- really anything.

For more specifics on how to write a decision engine, check out the decision engine documentation.

For everything else...

$ dharma faucet

Opens up a wizard for requesting free testnet ether from Dharma's faucet.

$ dharma init <path>

Creates an example decision engine at the given path

$ dharma wallet

Opens up a wizard for managing your funds, sending test ether to friends, and making loan repayments.

$ dharma authenticate <authToken>

Saves an auth token locally for future use. This allows you to solicit attestations from Dharma Labs Inc. and, in turn, access credit on the Dharma Network (see Overview onwards).


Loans in the Dharma Protocol can be thought of as miniature, borrower-initiated ICO's -- each loan is codified as a crowdfunding contract with attached metadata for loan terms, interest rates, and miscellaneous parameters. The smart contract serves as a vehicle for crowdfunding the loan, storing loan terms & parameters, distributing repayments to investors on a pro-rated basis according to each investor's individual contribution, and tracking borrower defaults and delinquencies.

When an investor contributes to a loan's crowdfund, they receive ERC20 tokens representing their ownership and rights to future cash flows in the loans. As such, loan tokens can be easily combined and repackaged into smart contracts representing virtually any tranched derivative debt instrument. Moreover, loan tokens are as transferrable as any other digital asset, meaning they can (in theory) be bought and sold on exchanges.

Credit risk assessment and identity verification is performed by trusted, centralized third parties known as Risk Assessment Attestors (RAAs). For a pre-defined fee that is codified into the loan contract, RAAs use whatever means are at their disposal to assess a borrower's creditworthiness and cryptographically sign a statement predicting the borrower's likelihood of default. In theory, market forces should gravitate loan volume towards those RAAs that produce better default risk predictions than others, given that a loan's performance can easily be audited ex post facto on chain. Dharma Labs Inc. currently acts as the sole RAA of the protocol, but, in the future, we plan on developing decentralized mechanisms for authenticating other RAAs into the system.

How It Works

Loans in the Dharma Protocol can be thought of as simple crowdfunding contracts with attached lending-specific metadata. In practice, loans aren't represented by individual smart contracts, but rather, for gas cost-saving reasons, stored within a meta-contract that manages loan mechanics in the entire Dharma Loan network. If a borrower wants a loan in the Dharma Protocol, they first must solicit a signed attestation from an RAA, at which point they create a loan request contract within the larger meta-contract with the following included:

  1. The RAA's signed attestation
  2. Their desired principal amount
  3. The terms of the loan
  4. Parameters around the timeline of the auction (i.e. the length of the auction in blocks)

Then, a fixed amount of time (denominated in blocks) is allotted for an auction period, in which investors can submit bids on the contract asserting their desired interest rates. Bidding consists of sending the contract a deposit equal to the maximum amount of ether the lender is willing to invest in the contract, and specifying the minimum interest rate the lender is willing to receive for that amount. Once the auction period is complete, the borrower has a pre-defined amount of time to either reject the given bids, or accept the bids by submitting to the contract any subset of the investors' bids whose total cash amount is equal to the desired principal + the RAA's pre-defined fee. The contract then assigns the loan an interest rate equal to the maximum of the accepted bids' interest rates and transfers the principal to the borrower. Lenders are then allowed to withdraw the remainder and / or entirety of their previous bids from the contract.

Borrower repayments are made directly to the smart contract, which automatically distributes payouts to the investors on a pro-rata basis according to each investor's individual contribution.

If a borrower defaults on a loan or is delinquent in their payments, their default will be visible on-chain, and their likelihood of receiving creditworthiness attestations from RAAs in the future will be greatly reduced.

Decision Engine Documentation

The Dharma CLI ingests ES7 Javascript classes that are required to expose a constructor and async method of the following format:

 constructor(web3) {
   this.web3 = web3;

 async decide(loan) {
     Decision logic goes here
         e.g. if (loan.defaultRisk < X && loan.principal > Y) { ... }

     If IS NOT bidding on loan, returns false

     If IS bidding on loan, returns object of the following form:
   return {
       Amount the investor wishes to bid on the loan, in Wei

           amount: <BigNumber>
     amount: new this.web3.BigNumber(2*(Math.pow(10, 18))),

       Minimum interest rate the investor is willing to accept for the loan.
       NOTE: since Solidity does not support floats, interest rates are
       expressed as 18 decimal BigNumbers (e.g. 23% interest = 0.23 * (Math.pow(10, 18)))

           minInterestRate: <BigNumber>
     minInterestRate: new this.web3.BigNumber(0.23*(Math.pow(10, 18)))

     Once the auction period is over, if enough bids have been cast, the
     borrower will choose whether or not to accept the given interest rate, and
     the unaccepted bids / remaining portions of accepted bids can be withdrawn
     by lenders.
The loan object exposes the following instance variables:



string - The uuid of the given loan request.



BigNumber - A BigNumber instance representing the principal requested in Wei



String - The Ethereum address of the borrower



Terms - The signed Terms object associated with the loan request (see Terms schema below)



String - The Ethereum address of the RAA attesting to the given loan.



BigNumber - A BigNumber instance representing the fee collected by the attestor in Wei. The fee is a fixed sum that is socialized across all investors in the loan (i.e. if an investor contributes 2/3 of the loan principal, they will pay 2/3 of the attestor fee)



BigNumber - A BigNumber instance representing the likelihood of default, as predicted by the RAA associated with this loan.


Returns An object with the following keys comprising the RAA's ECDSA signature of a stringified JSON object containing all of the above key-value pairs:

  r: <BigNumber>
  s: <BigNumber>
  v: <BigNumber>



BigNumber - A BigNumber instance representing the length of the loan's auction period, in blocks.



BigNumber - A BigNumber instance representing the length of the loan's review period, in blocks.

The Terms object exposes the following instance variables:



BigNumber - Current version of the terms schema



String - The time units in which the amortization periods of the loan are denominated in. Can be exclusively of types: daily, weekly, monthly, and yearly



BigNumber - The length of each payment period, in terms of periodType (i.e. if each payment period lasts 2 weeks, terms.periodType = 'weekly' and terms.periodLength = 2)



BigNumber - The number of payment periods in the entire loan term (i.e. if the entire term of the loan is 2 months and payments are expected every two weeks, terms.periodType = 'weekly', terms.periodLength = 2, and terms.termLength = 4)


NOTE: Currently, only non-compounded interest rates are supported by the CLI, so this variable is, by default, false.


Boolean - True or false indicating whether or not interest is compounded on the loan.