Jarvis is a voice commanding assistant service in Python 3.5+ It can recognize human speech, talk to user and execute basic commands.
- Opens a web page (e.g 'Jarvis open youtube')
- Play music in Youtube (e.g 'Jarvis play mozart')
- Increase/decrease the speakers master volume (also can set max/mute speakers volume) ** (e.g 'Jarvis volume up!')
- Opens libreoffice suite applications (calc, writer, impress) (e.g 'Jarvis open calc')
- Tells about something, by searching on the internet (e.g 'Jarvis tells me about oranges')
- Tells the weather for a place (e.g 'Jarvis tell_the_skills me the weather in London')
- Tells the current time and/or date (e.g 'Jarvis tell me time or date')
- Set an alarm (e.g 'Jarvis create a new alarm')
- Tells the internet speed (ping, uplink and downling) (e.g 'Jarvis tell_the_skills me the internet speed')
- Tells the internet availability (e.g 'Jarvis is the internet connection ok?')
- Tells the daily news (e.g 'Jarvis tell me today news')
- Spells a word (e.g 'Jarvis spell me the word animal')
- Creates a reminder (e.g 'Jarvis create a 10 minutes reminder')
- Opens linux applications (e.g 'Jarvis open bash/firefox')
- Tells everything it can do (e.g 'Jarvis tell me your skills or tell me what can you do')
- Tells the current location (e.g 'Jarvis tell me your current location')
- Tells how much memory consumes (e.g 'Jarvis tell me your memory consumption)
- Tells users commands history (e.g 'Jarvis tell me my history')
- Write/tell 'remember' and enable learning mode and add new responses on demand! (e.g 'Jarvis remember')
- Clear bash console (e.g 'Jarvis clear console')
- Has help command, which prints all the skills with their descriptions (e.g 'Jarvis help')
- Do basic calculations (e.g 'Jarvis (5 + 6) * 8' or 'Jarvis one plus one')
- Change settings on runtime (e.g 'Jarvis change settings')
- Asynchronous command execution & speech recognition and interpretation
- Supports two different user input modes (text or speech), user can write or speek in the mic.
- Answers in general questions (via call Wolfram API), e.g ('Jarvis tell me the highest building')
- Change input mode on run time, triggered by a phrase e.g 'Jarvis change settings')
- Easy voice-command customization
- Configurable assistant name (e.g 'Jarvis', 'Sofia', 'John' etc.) (change on run time supported)
- Log preview in console
- Vocal or/and text response
- Keeps commands history and learned skills in MongoDB.'
Jarvis assistant uses third party APIs for speech recognition,web information search, weather forecasting etc. All the following APIs have free no-commercial API calls. Subscribe to the following APIs in order to take FREE access KEYs.
- OpenWeatherMap: API for weather forecast.
- WolframAlpha: API for answer questions.
- IPSTACK: API for current location.
- Download the Jarvis repo localy:
git clone https://github.com/ggeop/Jarvis.git- Setup Jarvis and system dependencies:
bash setup.sh- Put the Keys in settings
NOTE: For better exprerience, before you start the application you can put the free KEYs in the settings.py
nano Jarvis/src/jarvis/jarvis/setting.py
- Start the assistant service:
bash run_jarvis.shYou can easily add a new skill in two steps.
- Create a new configurationin SKILLS in skills/registry.py
{
'enable': True,
'func': Skills.new_skill,
'tags': 'tag1, tag2',
'description': 'skill description..'
} - Create a new skill module in skills/collection
The skill extraction implement in a matrix of TF-IDF features for each skill.
In the following example he have a dimensional space with three skills.
The user input analyzed in this space and by using a similarity metric (e.g cosine) we find the most similar skill.
- PRs are welcome
☺️ - Try to follow PEP 8 guidelines and add comments!