node package manager

This is the command-line client for API. If you install it, you can do a few things with

Note that this is just the command-line tool; for calling something from your own code, try the SDK.


Copyright 2016

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.


To install this program, you need node.js and npm.

You can then just use this command to install the program:

npm install -g

This will make the command available on your system.


You can do various things with the command-line tool. At any time, you can get the up-to-date usage info by using this call: -h

That should output all the commands you can call. Here's more information on each one. create

Create a new agent from a JSON or CSON file. This will create a new agent based on a single agent that is the contents of the file. (See for what should go in that file.)

If the filename ends with ".json", it'll be treated as JSON, and if it ends with ".cson", it'll be treated as CSON.

The agent will be created on the server, and the new version (including timestamps and ID) will go to the output. If you use the -q option, only the new agent ID will be output. read

Fetches an agent from the server to output. Use the name or the ID of the agent. If the name has spaces, wrap it in quotes (""). Always returns JSON.

If you have more than one agent with the same name, this will give an error and show you all the IDs with that same name. update

Updates an agent by name or ID given the contents of the agent file. The file can be JSON or CSON, as with create, and the new agent will be sent to the output. delete

Deletes an agent by name or ID. There's no undo, so use this carefully! batch

Evaluate every line in this CSV file as an input set, and output the results as a CSV file including the inputs and outputs. You can use the agent name or ID.

The first line should be headers with the expected input names. Any columns in the CSV file that aren't used by the agent will be passed through to the output. So if you have an agent that takes input1 and input2 and outputs output1, and your input file looks like this:


Then the output will look like this:



You can use various options to change how the program behaves.

-k, --key

This is the API key that the command-line tool will use for the API. You can find your API key on the dashboard or in your settings.

It is required. If you use the command-line tool with the same key very often, you should consider putting it into a configuration file or environment variable (see below).

-r, --root

Root of the API server. Unless you are testing this particular program with a mock, don't change this value.

-o, --output

The output file to write to. Most of the commands have some output; this will put the output into this output file.

-q, --quiet

Output less data. Only kind of useful.

-b, --batch-size

When doing a bunch of evaluations with batch, this sets the number of evaluations to send at once. Defaults to 128, which is a lot of evaluations.

-c, --config-file

Path to the config file (see below).

-h, --help

Output help information. An abbreviated version of this document.

Configuration file

You can use a configuration file if you use the same arguments, like -k, over and over and over. It's a JSON file, by default in your home directory at $HOME/, but if you want you can move it somewhere else. It contains a JSON object with one property per option:

    "key": "your API key here"

Environment variables

You can also define environment variables for the default values of options. They are all prefixed with FUZZY_AI_, so these are the variables you can set: