Google have disabled AutoVoice on Google Assistant in some languages (English IS working, but some other languages are not). Please check here for more info and a temporary workaround.
Update: English is back up. Check here.
- Finished porting DialogFlow API to V2
- Fixed AutoVoice Recognize without a dialog not working on Android 10
- Started to update DialogFlow API to V2 which requires some actions by the user
All beta features are now public.
- Migrated from GCM to FCM
- Updated Google Play Services to their latest versions
- Re-introduced the option to use the app for free by watching a rewarded video ad
- Added option for offline recognition in the AutoVoice Recognize action
- Uses Services for Tasker plugin communication so it’s faster and more reliable
Beta features out for everyone!
3.0.9b (Beta) (2017-12-01)
- AutoVoice is now free to use with the Amazon Echo! That means that even the free lite version will work with all commands from the echo! 🙂
- AutoVoice will work in German, French, Italian, Spanish and Portuguese on the Google Assistant in the near future
- New AutoVoice for Google Assistant linking process that’s much easier and error free! Demo here.
- API.AI renamed to Dialogflow in the app
- Use AutoVoice as a real assistant on your device with full “Ok Google” support! Demo video here.
- Preparations for the Smart Home Amazon Echo skill. Will not work until approved. Demo video here.
- AutoVoice is now available on the Google Assistant on Android devices!
- Added new AutoVoice Natural Language Responses action in Tasker that allows you to set Natural Language command responses dynamically
- Added ability to edit all kinds of responses from the main app -> Natural Language -> Commands -> Responses
- Removed permission to access contacts for signing in to your Google Account
- Added ability to use Natural Language as the assistant when long-pressing the home button on your device
- Added reminder that you can set AutoVoice as your device’s assistant. Check out the how-to video.
- Added ability to change output contexts lifespans in natural language commands
- fixed bug where editing output contexts would reset their lifespans to 5
- Added ability to import natural language commands and types directly from the AutoVoice UI
- fixed issue where “Easy Commands” weren’t always being triggered when using the AutoVoice Test action in Tasker
- Added ability to export and import Natural Language commands and types
- Added %avsource variable in Natural Language condition in Tasker that lets you know where the command came from (Google Now, Assistant, etc)
- Fixed sorting of Natural Language commands by name
Beta Features now available for everyone! Some highlights:
- Support for Google Home and Amazon Echo. Get started here.
- Natural Language
- IFTTT integration with natural language
Support for Google Home, Amazon Echo and Natural Language is now available for everyone!
Check the beta releases below for more details on the changes
3.0.10b (Beta) (2017-01-30)
- Fixed deleting Assistant Commands
- Fixed variables in natural language commands not being shown after created through AutoVoice sometimes
- Fixed creating a new natural language command through Tasker bringing up the AutoVoice app
- Fixed crash when toggling the “Use for Assistant” switch in the Natural Langauge commands screen
- Made Natural Language commands generate AutoApps and IFTTT commands by ordering the parameters in alphabetical order
- Added video tutorial link under AutoVoice -> Personal Assistants
- Made creating a natural language voice command in tasker automatically select the command that was just created
- other smaller natural language fixes
- Added options to edit natural language types’ Reference Mode and Allow Other Values features
- Some bug fixes related to Natural Language
- Made some adjustments in preperation for the upcoming Amazon Alexa support
- Made commands received from “Note to Self” in Google Now and the Google Now XPosed API be able to be processed with Natural Language
- Added IFTTT support for natural voice commands. Read all about it here.
3.0.3b (Beta) (2016-12-28)
- Added in Natural Language action in Tasker:
- Reset Contexts: allows you to start a new voice command from scratch without having to answer to an unfinished command questions
- Minimum Confidence: set a level of confidence from which commands will “count”. If below this level commands won’t be considered valid.
- Ask Again: If set, will ask user to say voice command again with a prompt if minimum confidence has not been reached.
- Added ability to choose a custom user-created type when creating new natural voice command
- Added ability to create new custom type from within AutoVoice
- Added type editing screen for Natural Language
- Added ability to set prompts for parameters from AutoVoice
- Fixed natural language google now integration: will not cancel all searches anymore
- Updated two tutorials (here and here) so that they show creating natural language commands in AutoVoice instead of API.AI
- Google Home integration! The AutoVoice assistant is currently under review by Google but once it’s approved you’ll be able to say “Ok google, ask AutoVoice to ” and then any AutoVoice command you want and it’ll execute it on your devices!
- You also get Natural Voice Commands integration with Google Home! Because this takes a big toll on AutoVoice’s server you’ll need to subscribe to a monthly subscription to use this.
- Natural Language revamped: you can now create and edit natural language commands right from the app. Full video example here and tutorial here.
- AutoVoice commands made easy: creating a new AutoVoice profile is a breeze! Give it a list of possible commands, set some variables and off you go! Check out this video example.
- Natural Language processing! Convert loosely structured voice commands to very precise actions and variables. Example video here.
- Added Google Assistant integration workaround while the Assistant API is not public. Check out a tutorial here.
- Fixed bug on AutoVoice Test Command action
- Added notice about AutoVoice not working with Google Now from the lock screen
- added option to have Google Now integration run in the foreground. This helps with some phones *cough*SAMSUNG*cough* that aggressively kill services in the background to save on battery Let me know if this helps 🙂
- Remote voice! Send AutoVoice commands from Windows, Linux Mac or another Android device. Instructions here: http://joaoapps.com/autovoice/remote/
- offline recognition now working with built-in AutoVoice recognize action (still not working with Google Now)
- fixed google now integration for latest Google App
- You can now send AutoVoice commands to any device you own. More info here
2.0.9b (Beta) (2014-07-24)
- Added new “AutoVoice Google Results” Tasker condition that lets you react to various Google Search Results. For now you get Song Search, so you can directly access artist, album and song from the song search. This replaces the method I showed some days ago.
- Fixed another BT headset issue
2.0.8b (Beta) (2014-07-24)
- Fixed issues with BT headset button presses and voice recognition. Let me know if there’s still any issue with this.
- Fixed the “Double Tap to edit bug” for real now. 🙂 Edit your specific phrase in the AutoVoice settings
2.0.6b (Beta) (2014-07-24)
- Made it so that when you do a song search on Google Now (“what’s this song” command), the result will also be sent to AutoVoice, so that you can do something like in this video
2.0.3b (Beta) (2014-07-22)
- Added option in “AutoVoice Recognized” to make condition not cancel Google Search if used from Google Now
- Cleaned up language options in “AutoVoice Recognize” action. The only languages that show up now are supported languages. I also added the language codes for each language so that you can easily look them up.
- Fixed issue where some old AutoVoice profiles would not be matched
- Added option to remove custom phrase from Google Now commands, since some people were having “Double tap to edit” or some similar phrase added to all their commands. Find it in the “Advanced” screen in the Google Now Settings
- Added Google Now logs to try and find out why Google Now doesn’t work for some people.
- fixed crash for devices running Gingerbread or lower
- fixed issue with “Double tap to edit” on Google Now commands that some people were having
- to everyone that had their old AutoVoice profiles stop working, can you get in contact with me please? I couldn’t track down the problem yet.
- AutoVoice is now out for everyone 🙂 Check all the changes below in the beta versions. These are some highlights:
- Google Now commands can now be intercepted with AutoVoice via an Accessibility service! No Root required!
- AutoVoice now returns the spoken voice command in the same Task, so that you can react to it right there.
- Added Achievements. Learn AutoVoice in a fun way!
- Added %regexgroups() variable. This allows you to get ALL regex groups, not just named.
- Added third-party integration support.
- Fixed option to not show a popup when recognizing voice
1.5.37b (Beta) (2014-07-14)
- Added option to only match voice commands in Google Now, enabled by default. Can anyone with a MotoX confirm that this works with touchless controls?
- Added option to disable Google Now integration
- Added information over the state of the AutoVoice accessibility service
- Added more achievements
- Added “Invert” option in the “No Match” condition that allows you to make the profile trigger if any “AutoVoice Recognized” condition is matched
- Google Now commands can now be intercepted with AutoVoice via an Accessibility service! Enable the service in your Android settings->Accessibility->AutoVoice Google Now Integration and AutoVoice should intercept recognized commands in Google Now! Let me know if something’s not working right!
1.5.35b (Beta) (2014-07-09)
- Can now run voice commands with Google now, no root required! Simply say “Note to Self” followed by the voice command you want to execute and choose AutoVoice as the default handler!
- Added some more achievements
1.5.28b (Beta) (2014-05-29)
- Fixed default settings so that they “stick”. They were getting reset every time you entered the Default Settings config screen
- Since Google fixed the voice recognition popup, made that the default again. The notification method could conflict with Google’s launcher if it had background voice recognition active.
- lite version is now ad supported
- added support for achievements and leaderboards (no achievements yet)
1.5.25b (Beta) (2014-05-14)
- Added setting so that you can set default settings when launching AutoVoice from Google Now’s swipe up shortcut or directly from a Bluetooth Headset’s button. This lets you change the language you want to recognize in when launching from the Google Now shortcut for example.
- Fixed AutoVoice No Match in some situations.
- Added %regexgroups() variable. This allows you to get ALL regex groups (whether they are named or not, and independently of the level of nesting you’re using in the regex ) in the Task, so you can regex away to your heart’s desire 🙂 This means that, for example, the regex ((this) (is awesome)) will produce a %regexgroups() array like “this is awesome,this,is awesome”
- Added support for this XPosed Module to make Touchless controls work with AutoVoice. This works using the new “AutoVoice Ok Google Now” Tasker event.
- Removed options related to the Google Now API Xposed module, since it’s not working any more.
- Made screen not turn on if not using the popup for voice recognition
- Fixed bug related to devices on Android versions lower than 3.0
- Added warning when AutoVoice can’t start recognizing because the microphone is in use.
1.5.23b (Beta) (2014-05-02)
- Fixed bug where BT headset wouldn’t be used for voice recognition
- Fixed bug where sometimes voice results could not be correctly returned in the same task.
1.5.22b (Beta) (2014-04-29)
- Added third-party integration support. Use the following code to send voice commands to AutoVoice from another app (replace “myapp” with the name of your app; this will appear in the %avsource variable in Tasker):
ArrayList heard = getHeard();
Intent intent = new Intent("com.joaomgcd.autovoice.ACTION_AUTOVOICE_RECOGNIZED");
- Added timeout option in the “AutoVoice Recognize” Tasker action
- Fixed language bug (recognition language was not being changed)
- Other bugfixes
1.5.19b (Beta) (2014-04-18)
- Made the “AutoVoice Recognize” Tasker action return results in the same Task as it was called. This will allow you to do command chains much easier now, no longer needing to set Command IDs and whatnot 🙂 Please let me know if something’s not working with it.
- several small bugfixes
1.5.18b (Beta) (2014-04-17)
- Made all eligible AutoVoice conditions available as events. Please use those instead of the conditions if you can because they are much better performance wise.
1.5.15b (Beta) (2014-04-15)
- Changed the way the “AutoVoice Recognize” Tasker action works. Now the normal recognition prompt won’t be shown by default, but a notification will. This allows it to bring back the old AutoVoice behaviour to accept your command as soon as you finish talking. It also takes care of the bug that was happening if you selected the “No Popup” option before. It also allows the action to be used while the screen is off.
- Please let me know if this new feature works with your bluetooth headsets as before. I lost mine recently so I can’t test it right now. Thanks!
- Fixed bug with the “AutoVoice Current Noise” action where it would return a wrong value on some devices
- attempt to fix issue where AutoVoice continuous would stop if the microphone was used by another app.
- Lets you choose which audio stream to mute when using AutoVoice Continuous
- Added %avsource variable in the “AutoVoice No Match” condition
- Added new Tasker action ‘AutoVoice Curr Noise’ that allows you to get the current ambient noise. This allows you to set volumes based on the current ambient noise for example.
- Added language model option to the “AutoVoice Recognize” action. This allows you to get different kinds of results for the same spoken command.
- Added option to only filter Google Now searches to only Voice searches, text searches or both.
- Fixed bug where the No Matches condition would not return all recognized results
- Made AutoVoice compatible with Google Now!
- Made AutoVoice compatible with devices with no bluetooth or touchscreen
Production version released with all the Beta version changes included!
– Added way to perform replacements in your commands. For example, if you’re trying launch an app with AutoLaunch, and you say “Open Google plus”, you can now easily replace “plus” with “+”, so AutoLaunch can open the app directly
– Added command filter option to match any word in the filter, no matter what order you put it. You can also use double quotes to match multiple words. For example, if you have the command filter ‘”turn on” lights bedroom’, and you check this option, it will match all these voice commands: “turn on the lights in my bedroom”, “hey bedroom turn on your lights”, “I wish the lights in the bedroom would turn on”
– Added way to test your voice commands both from inside Tasker with the “AutoVoice Test Command” action and in the main app itself. Basically you write the command and AutoVoice pretends you said it and acts accordingly
– Set a name for your condition so you can more easily debug it with the AutoVoice logs
– No Match condition fixed for AutoVoice continuous
– Continuous and Ambient Noise monitors now have an option to use the LED on the phone while they’re on
– Set the prompt text that shows up on the voice recognition popup
– Intercepts long search button press on older phones
– Trigger words work for continuous and “normal” mode
**Ambient Noise Detection**
Now you can setup AutoVoice to trigger a Tasker profile based on Ambient Noise!
You can for instance create a baby monitor, where if the sound level is above a certain value, it will call your other phone. Or better yet, send it an AutoRemote message. 😉
– Added option to only trigger a command on continuous or normal mode. So, for instance, you can make a “Hello” command only be matchable when you’re using the normal popup mode so you don’t trigger it by accident on continuous
– Changed way trigger words work. Should work better now.
– No Matches condition can set command it and be activated depending on the last command id too
– No matches condition now returns variable array %avcomms
**Making Event Behaviour useful!**
Ever felt frustrated that you are constantly enabling Event Behaviour when there doesn’t seem to be a good use case for not having it enabled? Well now there is!
Now, if you disable Event Behaviour you set an “Exit Command”. On that profile, the normal command activates the Tasker profile, and the “Exit command” deactivates it.
This way you can for instance say “I’m home” and activate your home profile, and say “Bye” and deactivate it. No longer do you need 2 seperate profiles for these situations.
**Continuous mode Trigger Words**
Now you can set optional trigger words in AutoVoice conditions. When you set those trigger words, the profile will only become active if the Trigger Word was spoken in the prior 5 seconds (or any time you want). This can be useful for commands you don’t want to trigger accidentally all the time that have common words like “photo”, for example.
For example, you can say “Ok Tasker” and then “Photo”, and it’ll take a Photo, but it won’t take a photo if you don’t say “Ok Tasker” beforehand.
– if you have the latest Tasker beta, AutoVoice variables will show up in the variable list in the entry tasks of AutoVoice profiles
– Continuous Foreground mode. AutoVoice should not be killed by Android any more if on, so makes continuous mode more stable.
– Continuous toggle in the main app settings screen.
– Language option in Continuous, normal and shortcut: set the recognition language independently from the system recognition language
– New variables: %avcomms and %avcommsnofilter. Contain all results and not just the first
– Can do normal 1 time recognition without popup dialog
– Seperate toast setting for 1 time and continuous recognition
– Global variable %AVCONTINUOUS is set in tasker, to indicate if continuous service is on or off. Enable external access in Tasker to make this work!
– Added Google Now gesture interception