Furhat Remote API

Note: This feature is considered a public BETA and can change significantly based on user feedback!

Furhat Remote API is a way to connect and give commands to your Furhat robot from a program running on an external computer on the same network. A number of programming languages are supported, including Python, C#, JavaScript, Rust etc. (50+ languages, for full list see Swagger documentation)

If you want to to use it from Python, we recommend to use the Python PyPi library, which wraps the Remote API and makes it easier to install and use.

Setup

Run the server on the Robot

Download the Remote API skill file and load it to you Furhat robot. When the skill is started, the robot will start a Swagger Kotlin server listening on port 54321.

Testing the remote API using Postman

If you want, you can test the Remote API with the tool Postman:

  1. Download Postman,
  2. Download and open the Furhat Remote API yaml specification,
  3. Test the different requests (look for API documentation below).

Create a client for your specific programming language

In this step you will generate the code in your preferred programming language that will enable your program to communicate with Furhat.

  1. Download swagger or use their online editor,
  2. Paste the content from Furhat Client yaml file ,
  3. Click 'Generate Client' and select your language,
  4. Incorporate the client to your system and send requests to control the robot.

API Documentation

API Endpoints

All URIs are relative to http://IP of the robot:54321

Method HTTP request Description
furhatAttendPost POST /furhat/attend Attend a user/location
furhatFacePost POST /furhat/face Change the texture and/or model
furhatGesturePost POST /furhat/gesture Perform a gesture
furhatLedPost POST /furhat/led Change the colour of the LED strip
furhatListenGet GET /furhat/listen Make the robot listen, and get speech results
furhatListenStopPost POST /furhat/listen/stop Make the robot stop listening
furhatSayPost POST /furhat/say Make the robot speak
furhatSayStopPost POST /furhat/say/stop Make the robot stop talking
furhatVoicePost POST /furhat/voice Set the voice of the robot
furhatGesturesGet GET /furhat/gestures Get all gestures
furhatGet GET /furhat Test connection
furhatUsersGet GET /furhat/users Get current users
furhatVoicesGet GET /furhat/voices Get all the voices on the robot

Models

Authorization

All endpoints do not require authorization.

/furhat

GET
Summary:

Test connection

Description:

Used to verify if the server is running, return "hello world" upon success

Responses
Code Description
200 Status update

/furhat/gestures

GET
Summary:

Get all gestures

Description:

Returns a JSON array with all gestures on the system (names + duration).

Responses
Code Description Schema
200 A list of possible gestures [ Gesture ]

/furhat/voice

POST
Summary:

Set the voice of the robot

Description:

Sets the voice of the robot using the name of the voice, can be requested by doing a GET request on this endpoint.

Parameters
Name Located in Description Required Schema
name query The name of the voice Yes string
Responses
Code Description Schema
200 Successfull operation Status

/furhat/voices

GET
Summary:

Get all the voices on the robot

Description:

Returns a JSON array with voice names + languages.

Responses
Code Description Schema
200 Success [ Voice ]

/furhat/users

GET
Summary:

Get current users

Description:

Get all current users (max: 2). Returns a JSON array containg Users (Rotation, Location, id).

Responses
Code Description Schema
200 successful operation [ User ]

/furhat/say

POST
Summary:

Make the robot speak

Description:

Makes the robot speak by either using text, or a URL (linking to a.wav file). If generatelipsync=true, it uses a .pho file hosted on the same url, or generates phonemes by itself.

Parameters
Name Located in Description Required Schema
text query A string depicting a utterance the robot should say. No string
url query A url link to a audio file (.wav) No string
blocking query Whether to block execution before completion No boolean
lipsync query If a URL is provided, indicate if lipsync files should be generated/looked for. No boolean
abort query Stops the current speech of the robot. No boolean
Responses
Code Description Schema
200 successful operation Status

/furhat/attend

POST
Summary:

Attend a user/location

Description:

Provides 3 modes of attention. 1. Attend user based on enum (CLOSEST, OTHER or RANDOM) 2. Attend user based on it's id (can be retrieved by using /furhat/users) 3. Attend location based on coordinates (x,y,z)

Parameters
Name Located in Description Required Schema
user query Make furhat attend a user. Example 'CLOSEST' No [ ]
userid query Make furhat attend specified user No string
location query Make furhat attend location, usage: x,y,z. Example -20.0,-5.0,23.0 No string
Responses
Code Description Schema
200 Successfull operation Status

/furhat/face

POST
Summary:

Change the texture and/or model

Description:

Changes the texture or model, based on the model/textue name. Case sensitive. Names can be retrieved from the web interface

Parameters
Name Located in Description Required Schema
model query Change the model of the robot No string
texture query Change the texture of the robot No string
Responses
Code Description Schema
200 Successfull operation Status

/furhat/gesture

POST
Summary:

Perform a gesture

Description:

Performs a gesture based on 1. Gesture name (retrieve by GET request to /furhat/gestures) 2. Gesture definition, see example

Parameters
Name Located in Description Required Schema
name query The gesture to do No string
blocking query Whether to block execution before completion No boolean
definition body No GestureDefinition
Responses
Code Description Schema
200 Successfull operation Status
400 Parameters are wrong Status

/furhat/listen

GET
Summary:

Make the robot listen, and get speech results

Description:

Blocking call to get user speech input, language defaults to english_US. Language parameter can be used to provide a different language. Return values can be found in the Status object as message and can be: - User speech - SILENCE - INTERRUPTED - FAILED

Parameters
Name Located in Description Required Schema
language query The language to listen for, defaults to en-US No string
Responses
Code Description Schema
200 Successfull operation Status

/furhat/listen/stop

POST
Summary:

Make the robot stop listening

Description:

Aborts the listen

Responses
Code Description Schema
200 Successfull operation Status

/furhat/led

POST
Summary:

Change the colour of the LED strip

Description:

Changes the LED strip of the robot, colours can be between 0-255 (above 255 is changed to 255). Any parameter not provided defaults to 0.

Parameters
Name Located in Description Required Schema
red query The amount of red No integer
green query The amount of green No integer
blue query The amount of blue No integer
Responses
Code Description Schema
200 Successfull operation Status

/furhat/say/stop

POST
Summary:

Make the robot stop talking

Description:

Stops the current speech.

Responses
Code Description Schema
200 Successfull operation Status

Models

Status
Name Type Description Required
success boolean No
message string No
Gesture
Name Type Description Required
name string No
duration integer (double) No
User
Name Type Description Required
id string No
rotation Rotation No
location Location No
Location
Name Type Description Required
x double No
y double No
z double No
Voice
Name Type Description Required
name string No
language string No
Rotation
Name Type Description Required
x double No
y double No
z double No
GestureDefinition

The class name needs to be furhatos.gestures.Gesture otherwise it won't be parsed as a gesture. Examples can be found below.

Name Type Description Required
name string No
frames [ Frame ] No
class string No
GestureDefinition examples

All basic params are listed here:

    //All parameters have values between 0.0 and 1.0 (Except for the ones at the bottom).
    EXPR_ANGER
    EXPR_DISGUST
    EXPR_FEAR
    EXPR_SAD
    SMILE_CLOSED
    SMILE_OPEN
    SURPRISE
    BLINK_LEFT
    BLINK_RIGHT
    BROW_DOWN_LEFT
    BROW_DOWN_RIGHT
    BROW_IN_LEFT
    BROW_IN_RIGHT
    BROW_UP_LEFT
    BROW_UP_RIGHT

    EYE_SQUINT_LEFT
    EYE_SQUINT_RIGHT

    LOOK_DOWN
    LOOK_LEFT
    LOOK_RIGHT
    LOOK_UP

    PHONE_AAH
    PHONE_B_M_P
    PHONE_BIGAAH
    PHONE_CH_J_SH
    PHONE_D_S_T
    PHONE_EE
    PHONE_EH
    PHONE_F_V
    PHONE_I
    PHONE_K
    PHONE_N
    PHONE_OH
    PHONE_OOH_Q
    PHONE_R
    PHONE_TH
    PHONE_W

    LOOK_DOWN_LEFT
    LOOK_DOWN_RIGHT
    LOOK_LEFT_LEFT
    LOOK_LEFT_RIGHT
    LOOK_RIGHT_LEFT
    LOOK_RIGHT_RIGHT
    LOOK_UP_LEFT
    LOOK_UP_RIGHT

    //The following parameters have values in the range -50.0 to 50.0
    NECK_TILT
    NECK_PAN
    NECK_ROLL
    GAZE_PAN
    GAZE_TILT

A couple of gesture examples:

Built-in BigSmile

{
  "name":"BigSmile",
  "frames":[
    {
      "time":[0.32,0.64],
      "persist":false, <- Optional
      "params":{
        "BROW_UP_LEFT":1,
        "BROW_UP_RIGHT":1,
        "SMILE_OPEN":0.4,
        "SMILE_CLOSED":0.7
        }
    },
    {
      "time":[0.96],
      "persist":false, <- Optional
      "params":{
        "reset":true
        }
    }],
  "class":"furhatos.gestures.Gesture"
}

Custom gesture

{
  "frames": [
    {
      "time": [
        0.17, 1.0, 6.0
      ],
      "params": {
        "NECK_ROLL": 25.0,
        "NECK_PAN": -12.0,
        "NECK_TILT": -25.0
      }
    },
    {
        "time": [
            7.0
        ],
        "params": { 
            "reset": true
        }
    }
  ],
  "name": "Cool Thing",
  "class": "furhatos.gestures.Gesture"
}
Frame

A list of times can be provided, at those times the params will be executed.

Name Type Description Required
time [ double ] No
params BasicParam No
BasicParam

All supported parameters can be found here BasicParam.

Name Type Description Required
BasicParam object

Python Remote API

Description

To simplify the use of the Furhat Remote API from Python, there is a package on PyPi called furhat-remote-api.

Installation

You can install the package using pip:

pip install furhat-remote-api

(you may need to run pip with root permission: sudo pip install furhat-remote-api)

Usage

This shows how the different methods in the Remote API can be invoked from Python. 

from furhat_remote_api import FurhatRemoteAPI

# Create an instance of the FurhatRemoteAPI class, providing the address of the robot or the SDK running the virtual robot
furhat = FurhatRemoteAPI("localhost")

# Get the voices on the robot
voices = furhat.get_voices()

# Set the voice of the robot
furhat.set_voice(name='Matthew')

# Say "Hi there!"
furhat.say(text="Hi there!")

# Play an audio file (with lipsync automatically added) 
furhat.say(url="https://www2.cs.uic.edu/~i101/SoundFiles/gettysburg10.wav", lipsync=True)

# Listen to user speech and return ASR result
result = furhat.listen()

# Perform a named gesture
furhat.gesture(name="BrowRaise")

# Perform a custom gesture
furhat.gesture(definition={
    "frames": [
        {
            "time": [
                0.33
            ],
            "params": {
                "BLINK_LEFT": 1.0
            }
        },
        {
            "time": [
                0.67
            ],
            "params": {
                "reset": True
            }
        }
    ],
    "class": "furhatos.gestures.Gesture"
    })

# Get the users detected by the robot 
users = furhat.get_users()

# Attend the user closest to the robot
furhat.attend(user="CLOSEST")

# Attend a user with a specific id
furhat.attend(userid="virtual-user-1")

# Attend a specific location (x,y,z)
furhat.attend(location="0.0,0.2,1.0")

# Set the LED lights
furhat.set_led(red=200, green=50, blue=50)