Releases: WhiskerWeirdo/BanterBrain-Buddy
BanterBrainBuddy 1.0.7 Dont say any bad words
BanterBrain Buddy v1.0.7 feature release
This program requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
Please report all bugs to the Discord or here on Github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V1.0.7 RELEASE
New:
- Allow the ElevenLabs 2.5 Turbo API at 50% the cost of the default one https://elevenlabs.io/blog/introducing-turbo-v2-5
- You can now filter bad words in Twitch messages and triggers! If a bad word is found, the message is completely ignored by BBB
RELEASE V1.0.6 RELEASE
Fixed:
- Not saving Streamer name on exit #74
- Not saving LLM Language setting on exit #75
- Authorization to twitch failing due to ClientID not being correct in settings file #78
New:
- Allowing fully custom Twitch and intermediary LLM messages.
Note: There is no protection to what you enter or remove there! Be careful!
RELEASE V1.0.5 RELEASE
- Now supports OpenAI's ChatGPT 4 Omni Mini model.
RELEASE V1.0.4 RELEASE
Fixed:
- Check for Windows Native STT recognizer installed or not. Feedback when not #73
- Subscriber and Founder badge both recognized as issubscriber in new library. #71
- Crash on first install fix
New:
- You can now also select German and Spanish Twitch and LLM intermediary messages
RELEASE V1.0.3 RELEASE
New:
- shows if there is a version update in the bottom left
- Started working on multilingual Twitch messages (not yet finished)
- Recording button turns red when recording. Main textfield also shows when stopoing recording.
- Small text in hotkey setting to mention the difference between the hotkey (hold to record) and the button (toggle)
Fixed:
- Twitch channel command should not hang on subscriber chat commands
Not yet fixed:
- Native STT non English
Known issue:
- Viewers using their "First" badge while being subsribers are not correctly identified as subscribers
RELEASE V1.0.2 RELEASE
New:
- Spanish added to OpenAI/Whisper TTS voices
- ElevenLabs now defaults to the higher quality and multilingual v2 API speech model
Note: this does not fix the English used in some prompts that are embedded in BBB, which can generate some response oddities from the used LLM. This will be fixed in a later release.
RELEASE V1.0.1 RELEASE
New:
- Talking rate for TTS Voices that support it #62
- Talking volume for TTS Voices #53
- Talking pitch for TTS Voices that support it
- Version control check at startup
Supported TTS Voices features:
Azure Cognitive Services: pitch, volume, rate
Windows native: pitch, rate, volume
OpenAI Whisper: volume, rate
ElevenLabs: volume
RELEASE V1.0.0 RELEASE
Fixed:
- Some small logic errors for Elevenlabs fixed when loading twice in personas
- Some small logic errors fixed when switching personas and saving from dialog window
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error whenn the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.4 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
License
This project uses the GNU General Public License v3.0 as you can read here. The TLDR is:
You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions.
About
BanterBrain Buddy is a Windows .Net based Speech-To-Text to LLM to Text-To-Speech program for general entertainment or as a streaming companion.
The goal is to provide local PC or streaming entertainment by talking to an AI and hearing the responses back, based on a role you can set for the AI. The goa is to support both local only-resources or API services.
For the streaming side of things, integration with Twitch to respond to chat commands and stream events is planned.
Example: https://www.youtube.com/watch?v=TawapT1WEEo
Social
Come talk or chat at https://discord.banterbrain.tv
Build instructions
[todo]
tldr: use Visual studio 2022 with .net 8.
Resources for local and API
- for most API-services you need a paid account using a creditcard! (OpenAI, Azure, Google, Deepgram, etc.)
- local based LLM's like Ollama take significant resources; expect to need at least 8 GB memory for a basic model and a solid GPU if you want fast responses
BanterBrainBuddy 1.0.6
Note: Get a notification that this might be a virus? Try the .zip version!
BanterBrain Buddy v1.0.6 bugfix/feature release
This program requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
Please report all bugs to the Discord or here on Github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V1.0.6 RELEASE
Fixed:
- Not saving Streamer name on exit #74
- Not saving LLM Language setting on exit #75
- Authorization to twitch failing due to ClientID not being correct in settings file #78
New:
- Allowing fully custom Twitch and intermediary LLM messages.
Note: There is no protection to what you enter or remove there! Be careful!
RELEASE V1.0.5 RELEASE
- Now supports OpenAI's ChatGPT 4 Omni Mini model.
RELEASE V1.0.4 RELEASE
Fixed:
- Check for Windows Native STT recognizer installed or not. Feedback when not #73
- Subscriber and Founder badge both recognized as issubscriber in new library. #71
- Crash on first install fix
New:
- You can now also select German and Spanish Twitch and LLM intermediary messages
RELEASE V1.0.3 RELEASE
New:
- shows if there is a version update in the bottom left
- Started working on multilingual Twitch messages (not yet finished)
- Recording button turns red when recording. Main textfield also shows when stopoing recording.
- Small text in hotkey setting to mention the difference between the hotkey (hold to record) and the button (toggle)
Fixed:
- Twitch channel command should not hang on subscriber chat commands
Not yet fixed:
- Native STT non English
Known issue:
- Viewers using their "First" badge while being subsribers are not correctly identified as subscribers
RELEASE V1.0.2 RELEASE
New:
- Spanish added to OpenAI/Whisper TTS voices
- ElevenLabs now defaults to the higher quality and multilingual v2 API speech model
Note: this does not fix the English used in some prompts that are embedded in BBB, which can generate some response oddities from the used LLM. This will be fixed in a later release.
RELEASE V1.0.1 RELEASE
New:
- Talking rate for TTS Voices that support it #62
- Talking volume for TTS Voices #53
- Talking pitch for TTS Voices that support it
- Version control check at startup
Supported TTS Voices features:
Azure Cognitive Services: pitch, volume, rate
Windows native: pitch, rate, volume
OpenAI Whisper: volume, rate
ElevenLabs: volume
RELEASE V1.0.0 RELEASE
Fixed:
- Some small logic errors for Elevenlabs fixed when loading twice in personas
- Some small logic errors fixed when switching personas and saving from dialog window
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error whenn the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.4 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
License
This project uses the GNU General Public License v3.0 as you can read here. The TLDR is:
You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions.
About
BanterBrain Buddy is a Windows .Net based Speech-To-Text to LLM to Text-To-Speech program for general entertainment or as a streaming companion.
The goal is to provide local PC or streaming entertainment by talking to an AI and hearing the responses back, based on a role you can set for the AI. The goa is to support both local only-resources or API services.
For the streaming side of things, integration with Twitch to respond to chat commands and stream events is planned.
Example: https://www.youtube.com/watch?v=TawapT1WEEo
Social
Come talk or chat at https://discord.banterbrain.tv
Build instructions
[todo]
tldr: use Visual studio 2022 with .net 8.
Resources for local and API
- for most API-services you need a paid account using a creditcard! (OpenAI, Azure, Google, Deepgram, etc.)
- local based LLM's like Ollama take significant resources; expect to need at least 8 GB memory for a basic model and a solid GPU if you want fast responses
BanterBrain Buddy 1.0.5 GPT 4o mini
NOTE: Microsoft Defender seems to think this contains a virus. This for all appearances seems to be a false positive set off by updating a library. I have submitted this as a false positive to Microsoft, but can do nothing about this myself :(
Virusscanners results: https://www.virustotal.com/gui/file/64a183a284ed188a3e40f7f161d3cfa24d3adc8b0d0da60f95cdd556e0f68293
I created an alternative downloadable file using other installer settings which do not trigger Microsoft Defender: https://github.com/WhiskerWeirdo/BanterBrain-Buddy/releases/download/1.0.5_release/BanterBrainInstaller-1.0.5-alternative.msi
BanterBrain Buddy v1.0.5 feature release
This program requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
Please report all bugs to the Discord or here on Github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V1.0.5 RELEASE
- Now supports OpenAI's ChatGPT 4 Omni Mini model.
RELEASE V1.0.4 RELEASE
Fixed:
- Check for Windows Native STT recognizer installed or not. Feedback when not #73
- Subscriber and Founder badge both recognized as issubscriber in new library. #71
- Crash on first install fix
New:
- You can now also select German and Spanish Twitch and LLM intermediary messages
RELEASE V1.0.3 RELEASE
New:
- shows if there is a version update in the bottom left
- Started working on multilingual Twitch messages (not yet finished)
- Recording button turns red when recording. Main textfield also shows when stopoing recording.
- Small text in hotkey setting to mention the difference between the hotkey (hold to record) and the button (toggle)
Fixed:
- Twitch channel command should not hang on subscriber chat commands
Not yet fixed:
- Native STT non English
Known issue:
- Viewers using their "First" badge while being subsribers are not correctly identified as subscribers
RELEASE V1.0.2 RELEASE
New:
- Spanish added to OpenAI/Whisper TTS voices
- ElevenLabs now defaults to the higher quality and multilingual v2 API speech model
Note: this does not fix the English used in some prompts that are embedded in BBB, which can generate some response oddities from the used LLM. This will be fixed in a later release.
RELEASE V1.0.1 RELEASE
New:
- Talking rate for TTS Voices that support it #62
- Talking volume for TTS Voices #53
- Talking pitch for TTS Voices that support it
- Version control check at startup
Supported TTS Voices features:
Azure Cognitive Services: pitch, volume, rate
Windows native: pitch, rate, volume
OpenAI Whisper: volume, rate
ElevenLabs: volume
RELEASE V1.0.0 RELEASE
Fixed:
- Some small logic errors for Elevenlabs fixed when loading twice in personas
- Some small logic errors fixed when switching personas and saving from dialog window
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error whenn the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.4 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
License
This project uses the GNU General Public License v3.0 as you can read here. The TLDR is:
You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions.
About
BanterBrain Buddy is a Windows .Net based Speech-To-Text to LLM to Text-To-Speech program for general entertainment or as a streaming companion.
The goal is to provide local PC or streaming entertainment by talking to an AI and hearing the responses back, based on a role you can set for the AI. The goa is to support both local only-resources or API services.
For the streaming side of things, integration with Twitch to respond to chat commands and stream events is planned.
Example: https://www.youtube.com/watch?v=TawapT1WEEo
Social
Come talk or chat at https://discord.banterbrain.tv
Build instructions
[todo]
tldr: use Visual studio 2022 with .net 8.
Resources for local and API
- for most API-services you need a paid account using a creditcard! (OpenAI, Azure, Google, Deepgram, etc.)
- local based LLM's like Ollama take significant resources; expect to need at least 8 GB memory for a basic model and a solid GPU if you want fast responses
1.0.4 bugfix release
BanterBrain Buddy v1.0.4 bugfix
This program requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
Please report all bugs to the Discord or here on Github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V1.0.4 RELEASE
Fixed:
- Check for Windows Native STT recognizer installed or not. Feedback when not #73
- Subscriber and Founder badge both recognized as issubscriber in new library. #71
- Crash on first install fix
New:
- You can now also select German and Spanish Twitch and LLM intermediary messages
RELEASE V1.0.3 RELEASE
New:
- shows if there is a version update in the bottom left
- Started working on multilingual Twitch messages (not yet finished)
- Recording button turns red when recording. Main textfield also shows when stopoing recording.
- Small text in hotkey setting to mention the difference between the hotkey (hold to record) and the button (toggle)
Fixed:
- Twitch channel command should not hang on subscriber chat commands
Not yet fixed:
- Native STT non English
Known issue:
- Viewers using their "First" badge while being subsribers are not correctly identified as subscribers
RELEASE V1.0.2 RELEASE
New:
- Spanish added to OpenAI/Whisper TTS voices
- ElevenLabs now defaults to the higher quality and multilingual v2 API speech model
Note: this does not fix the English used in some prompts that are embedded in BBB, which can generate some response oddities from the used LLM. This will be fixed in a later release.
RELEASE V1.0.1 RELEASE
New:
- Talking rate for TTS Voices that support it #62
- Talking volume for TTS Voices #53
- Talking pitch for TTS Voices that support it
- Version control check at startup
Supported TTS Voices features:
Azure Cognitive Services: pitch, volume, rate
Windows native: pitch, rate, volume
OpenAI Whisper: volume, rate
ElevenLabs: volume
RELEASE V1.0.0 RELEASE
Fixed:
- Some small logic errors for Elevenlabs fixed when loading twice in personas
- Some small logic errors fixed when switching personas and saving from dialog window
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error whenn the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.4 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
License
This project uses the GNU General Public License v3.0 as you can read here. The TLDR is:
You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions.
About
BanterBrain Buddy is a Windows .Net based Speech-To-Text to LLM to Text-To-Speech program for general entertainment or as a streaming companion.
The goal is to provide local PC or streaming entertainment by talking to an AI and hearing the responses back, based on a role you can set for the AI. The goa is to support both local only-resources or API services.
For the streaming side of things, integration with Twitch to respond to chat commands and stream events is planned.
Example: https://www.youtube.com/watch?v=TawapT1WEEo
Social
Come talk or chat at https://discord.banterbrain.tv
Build instructions
[todo]
tldr: use Visual studio 2022 with .net 8.
Resources for local and API
- for most API-services you need a paid account using a creditcard! (OpenAI, Azure, Google, Deepgram, etc.)
- local based LLM's like Ollama take significant resources; expect to need at least 8 GB memory for a basic model and a solid GPU if you want fast responses
1.0.3 bugfix and version screen
BanterBrain Buddy v1.0.3 bugfix and version screen
This program requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
Please report all bugs to the Discord or here on Github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V1.0.3 RELEASE
New:
- shows if there is a version update in the bottom left
- Started working on multilingual Twitch messages (not yet finished)
- Recording button turns red when recording. Main textfield also shows when stopoing recording.
- Small text in hotkey setting to mention the difference between the hotkey (hold to record) and the button (toggle)
Fixed:
- Twitch channel command should not hang on subscriber chat commands
Not yet fixed:
- Native STT non English
Known issue:
- Viewers using their "First" badge while being subsribers are not correctly identified as subscribers
RELEASE V1.0.2 RELEASE
New:
- Spanish added to OpenAI/Whisper TTS voices
- ElevenLabs now defaults to the higher quality and multilingual v2 API speech model
Note: this does not fix the English used in some prompts that are embedded in BBB, which can generate some response oddities from the used LLM. This will be fixed in a later release.
RELEASE V1.0.1 RELEASE
New:
- Talking rate for TTS Voices that support it #62
- Talking volume for TTS Voices #53
- Talking pitch for TTS Voices that support it
- Version control check at startup
Supported TTS Voices features:
Azure Cognitive Services: pitch, volume, rate
Windows native: pitch, rate, volume
OpenAI Whisper: volume, rate
ElevenLabs: volume
RELEASE V1.0.0 RELEASE
Fixed:
- Some small logic errors for Elevenlabs fixed when loading twice in personas
- Some small logic errors fixed when switching personas and saving from dialog window
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error whenn the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.3 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
License
This project uses the GNU General Public License v3.0 as you can read here. The TLDR is:
You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions.
About
BanterBrain Buddy is a Windows .Net based Speech-To-Text to LLM to Text-To-Speech program for general entertainment or as a streaming companion.
The goal is to provide local PC or streaming entertainment by talking to an AI and hearing the responses back, based on a role you can set for the AI. The goa is to support both local only-resources or API services.
For the streaming side of things, integration with Twitch to respond to chat commands and stream events is planned.
Example: https://www.youtube.com/watch?v=TawapT1WEEo
Social
Come talk or chat at https://discord.banterbrain.tv
Build instructions
[todo]
tldr: use Visual studio 2022 with .net 8.
Resources for local and API
- for most API-services you need a paid account using a creditcard! (OpenAI, Azure, Google, Deepgram, etc.)
- local based LLM's like Ollama take significant resources; expect to need at least 8 GB memory for a basic model and a solid GPU if you want fast responses
Version 1.0.2 - TTS improvements
BanterBrain Buddy v1.0.2 TTS improvements
This program requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
Please report all bugs to the Discord or here on Github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V1.0.2 RELEASE
New:
- Spanish added to OpenAI/Whisper TTS voices
- ElevenLabs now defaults to the higher quality and multilingual v2 API speech model
Note: this does not fix the English used in some prompts that are embedded in BBB, which can generate some response oddities from the used LLM. This will be fixed in a later release.
RELEASE V1.0.1 RELEASE
New:
- Talking rate for TTS Voices that support it #62
- Talking volume for TTS Voices #53
- Talking pitch for TTS Voices that support it
- Version control check at startup
Supported TTS Voices features:
Azure Cognitive Services: pitch, volume, rate
Windows native: pitch, rate, volume
OpenAI Whisper: volume, rate
ElevenLabs: volume
RELEASE V1.0.0 RELEASE
Fixed:
- Some small logic errors for Elevenlabs fixed when loading twice in personas
- Some small logic errors fixed when switching personas and saving from dialog window
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error whenn the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.3 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
License
This project uses the GNU General Public License v3.0 as you can read here. The TLDR is:
You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions.
About
BanterBrain Buddy is a Windows .Net based Speech-To-Text to LLM to Text-To-Speech program for general entertainment or as a streaming companion.
The goal is to provide local PC or streaming entertainment by talking to an AI and hearing the responses back, based on a role you can set for the AI. The goa is to support both local only-resources or API services.
For the streaming side of things, integration with Twitch to respond to chat commands and stream events is planned.
Example: https://www.youtube.com/watch?v=TawapT1WEEo
Social
Come talk or chat at https://discord.banterbrain.tv
Build instructions
[todo]
tldr: use Visual studio 2022 with .net 8.
Resources for local and API
- for most API-services you need a paid account using a creditcard! (OpenAI, Azure, Google, Deepgram, etc.)
- local based LLM's like Ollama take significant resources; expect to need at least 8 GB memory for a basic model and a solid GPU if you want fast responses
Version 1.0.1 - TTS helium edition!
BanterBrain Buddy v1.0.1 helium edition
This program requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
Please report all bugs to the Discord or here on Github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V1.0.1 RELEASE
New:
- Talking rate for TTS Voices that support it #62
- Talking volume for TTS Voices #53
- Talking pitch for TTS Voices that support it
- Version control check at startup
Supported TTS Voices features:
Azure Cognitive Services: pitch, volume, rate
Windows native: pitch, rate, volume
OpenAI Whisper: volume, rate
ElevenLabs: volume
RELEASE V1.0.0 RELEASE
Fixed:
- Some small logic errors for Elevenlabs fixed when loading twice in personas
- Some small logic errors fixed when switching personas and saving from dialog window
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error whenn the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.3 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
License
This project uses the GNU General Public License v3.0 as you can read here. The TLDR is:
You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions.
About
BanterBrain Buddy is a Windows .Net based Speech-To-Text to LLM to Text-To-Speech program for general entertainment or as a streaming companion.
The goal is to provide local PC or streaming entertainment by talking to an AI and hearing the responses back, based on a role you can set for the AI. The goa is to support both local only-resources or API services.
For the streaming side of things, integration with Twitch to respond to chat commands and stream events is planned.
Example: https://www.youtube.com/watch?v=TawapT1WEEo
Social
Come talk or chat at https://discord.banterbrain.tv
Build instructions
[todo]
tldr: use Visual studio 2022 with .net 8.
Resources for local and API
- for most API-services you need a paid account using a creditcard! (OpenAI, Azure, Google, Deepgram, etc.)
- local based LLM's like Ollama take significant resources; expect to need at least 8 GB memory for a basic model and a solid GPU if you want fast responses
BanterBrain Buddy V1.0.0
v1.0.0 release of BanterBrain Buddy.
This program requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
Please report all bugs to the Discord or here on Github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V1.0.0 RELEASE
Fixed:
- Some small logic errors for Elevenlabs fixed when loading twice in personas
- Some small logic errors fixed when switching personas and saving from dialog window
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you @max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error whenn the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.3 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
License
This project uses the GNU General Public License v3.0 as you can read here. The TLDR is:
You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & install instructions.
About
BanterBrain Buddy is a Windows .Net based Speech-To-Text to LLM to Text-To-Speech program for general entertainment or as a streaming companion.
The goal is to provide local PC or streaming entertainment by talking to an AI and hearing the responses back, based on a role you can set for the AI. The goa is to support both local only-resources or API services.
For the streaming side of things, integration with Twitch to respond to chat commands and stream events is planned.
Example: https://www.youtube.com/watch?v=TawapT1WEEo
Social
Come talk or chat at https://discord.banterbrain.tv
Build instructions
[todo]
tldr: use Visual studio 2022 with .net 8.
Resources for local and API
- for most API-services you need a paid account using a creditcard! (OpenAI, Azure, Google, Deepgram, etc.)
- local based LLM's like Ollama take significant resources; expect to need at least 8 GB memory for a basic model and a solid GPU if you want fast responses
0.1.8-BETA
v0.1.8-BETA release of BanterBrain Buddy.
This BETA requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
This is a beta, please report all bugs to the discord or here on github.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
- ElevenLabs API check (preloading voices) can sometimes timeout after 15 seconds. This makes the first time you use it, probably quite a bit slower when editing persona's, but that's only until it works. This cannot really be fixed on my side.
RELEASE V0.1.8 BETA
Fixed:
- #57 now uses a broadcaster account to verify channel issues while a bot account can be used for talking in chat. The bot account is optional.
- #55 now throws an error when there is an audio device selected that's active but unusable with Azure Speech
- #58 logfile is now being created in appdata, The same directory where some application settings files are being written to. Also some more error handling incase there's issues writing.
RELEASE V0.1.7 BETA
Fixed:
- Streamer listbox STT was not cleared when coming back from settings #59
- Chat command cooldown timer for Twitch started at the beginning, not at the end of TTS #54
Need feedback:
- Lower delay between notification sound and saying response in TTS #61
- Log file not being created, added extra notification and testing #58
RELEASE V0.1.6 BETA
Thank you max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- Save dialog when you edit the default persona should not popup now all the time
not fixed:
- Added additional logging to eventsub subscription errors.
RELEASE V0.1.5 BETA
Thank you max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #50 by adding a small delay and a check for ratelimits
- (hopefully) Azure voice quality should be improved
Added:
- Ability to post the chat command cooldown being over in Twitch chat.
RELEASE V0.1.4 BETA
Thank you max aka Dadflaps#1337 for your bug reports and feedback
fixed:
- #45 fixed; due to logic error when the authorized API key and monitoring channel were different people
- #46 fixed; in combination with autostart checked, after the settings window Twitch would disconnect
- #49 fixed; username of broadcaster was not used in streamer/local STT recordings
- #47 fixed; setting was not saved when opening Settings panel.
Improvement:
- Switching windows now saves settings more often to prevent setting-data loss
- Main window now calls "Persona" "Streamer Persona" to make it more obvious that's only used for the Speech-To-Text/streamer.
RELEASE V0.1.3 BETA
- Enabling Ollama in the settings screen while its not running now handled gracefully instead of crashing. Testing Ollama now also handles issues more gracefully instead of crashing.
RELEASE V0.1.2 BETA
- Elevenlabs speed optimization. It only needs to load the voices when you are busy editing persona's that use Elevenlabs and not in other events.
RELEASE V0.1.1 BETA
- Fixes and validations for invalid or empty API keys so that the program does not crash. It fails gracefully and tells you the key is invalid.
RELEASE V0.1.0 BETA
- many fixes for text input issues, like saving on the correct time or not allowing empty fields
- Twitch OAUTH fix to allow configurable redirect if default port is occupied
- Elevenlabs optimizations
- Fixes to PTT button
- Fixes to Hotkeys
- Added timestamps to Main text window
- many code hygiene fixes
- Ollama roletext fix & now Ollama supports and remembers previous content
- Optimization of API verifications
- added more logging to help debugging
- Added a HELP setting that opens the logfile directory
- New Elevenlabs library
RELEASE V0.0.7 ALPHA 1
- Twitch can set if you want full response or just the answer generated by the bot
- Twitch can post the spoken bot respose into Twitch chat
- Supports multilingual Speech-To-Text for the AI bots to reply to.
RELEASE V0.0.7 ALPHA 1
- Allow an alert sound for Twitch events
- Speed up for Azure and Elevenlabs API's
RELEASE V0.0.7
- Supports local LLM Ollama
For more information see: wiki/How-do-I#use-ollama-the-local-llm
RELEASE V0.0.6-ALPHA3
- At first run a default STT and TTS provider is loaded and selected to prevent errors
- Twitch Trigger events now use the correct persona selected to respond
- Twitch personas now get saved when changed from the dropdown box
RELEASE V0.0.6-ALPHA2
- Fixed startup on crash about unset hotkey but also assigned a default
- Saving settings files (personas and settings) now in %appdata% directory
- Saving .wav file in %tmp% directory
- if no default microhphone in config file, default is selected
- if no default speaker in config file, default is selected
RELEASE V0.0.6-ALPHA1
- Implemented ElevenLabs TTS
- Fixed a bug for Twitch connection
RELEASE V0.0.5-ALPHA.4
- Help -> github now points to the wiki
- Wav files are unique and now get placed in a tmp directory and deleted when STT is done
- GPT temperature now is functional and can be used to create more unique and diverse responses
- Persona's other than default can be deleted. Default can still be edited however
- Native TTS/STT fully moved to its correct class
RELEASE V0.0.5-ALPHA.3
- Added STT hotkey
- Persona saved on settings hard close
RELEASE V0.0.5-ALPHA.2
- Added OpenAI Whisper Speech-to-Text cos why not.
RELEASE V0.0.5-ALPHA.1
- Twitch should not disconnect if you enter or close the settings screen
- Native TTS should work a lot better and should not throw an error of file in use
RELEASE V0.0.5-ALPHA
- Supports Windows Native Text-to-Speech and Speech-to-Text
- Supports Azure SpeechServices Text-to-Speech and Speec-to-Text (PAID SERVICE)
- Supports OpenAI ChatGPT 3.5-turbo API (PAID SERVICE)
- Supports Twitch integration using API and EventSub and triggers to: Chat command, Cheers (bits), Subscription, Resubscription, Gift subscription and custom channel point redemptions.
BanterBrain Buddy 0.0.7-ALPHA
v0.0.7-ALPHA.2 release of BanterBrain Buddy.
This ALPHA requires a valid OpenAI ChatGPT API key to operate or a local Ollama installation.
This being an ALPHA do not expect anything to work out of the box or easily. It's a public release to showcase the promise that BBB can have for the streaming community. Expect errors and problems and it's suggested to not use this release actively on Twitch for anything else than testing purposes.
Requisites to run
This release depends on .NET Runtime Desktop 8. After installation and trying to run BBB you will be asked to download and install this from the Microsoft website if you do not have it installed yet.
You can also download Runtime Desktop here: Windows Runtime Desktop 8.0.4
KNOWN ISSUES
- Plugging in/out audio devices while BBB runs can have an unforseen effect on your input/output and throw errors
- No ability to check for Twitch followers
RELEASE V0.0.7 ALPHA 1
- Twitch can set if you want full response or just the answer generated by the bot
- Twitch can post the spoken bot respose into Twitch chat
- Supports multilingual Speech-To-Text for the AI bots to reply to.
RELEASE V0.0.7 ALPHA 1
- Allow an alert sound for Twitch events
- Speed up for Azure and Elevenlabs API's
RELEASE V0.0.7
- Supports local LLM Ollama
For more information see: wiki/How-do-I#use-ollama-the-local-llm
RELEASE V0.0.6-ALPHA3
- At first run a default STT and TTS provider is loaded and selected to prevent errors
- Twitch Trigger events now use the correct persona selected to respond
- Twitch personas now get saved when changed from the dropdown box
RELEASE V0.0.6-ALPHA2
- Fixed startup on crash about unset hotkey but also assigned a default
- Saving settings files (personas and settings) now in %appdata% directory
- Saving .wav file in %tmp% directory
- if no default microhphone in config file, default is selected
- if no default speaker in config file, default is selected
RELEASE V0.0.6-ALPHA1
- Implemented ElevenLabs TTS
- Fixed a bug for Twitch connection
RELEASE V0.0.5-ALPHA.4
- Help -> github now points to the wiki
- Wav files are unique and now get placed in a tmp directory and deleted when STT is done
- GPT temperature now is functional and can be used to create more unique and diverse responses
- Persona's other than default can be deleted. Default can still be edited however
- Native TTS/STT fully moved to its correct class
RELEASE V0.0.5-ALPHA.3
- Added STT hotkey
- Persona saved on settings hard close
RELEASE V0.0.5-ALPHA.2
- Added OpenAI Whisper Speech-to-Text cos why not.
RELEASE V0.0.5-ALPHA.1
- Twitch should not disconnect if you enter or close the settings screen
- Native TTS should work a lot better and should not throw an error of file in use
RELEASE V0.0.5-ALPHA
- Supports Windows Native Text-to-Speech and Speech-to-Text
- Supports Azure SpeechServices Text-to-Speech and Speec-to-Text (PAID SERVICE)
- Supports OpenAI ChatGPT 3.5-turbo API (PAID SERVICE)
- Supports Twitch integration using API and EventSub and triggers to: Chat command, Cheers (bits), Subscription, Resubscription, Gift subscription and custom channel point redemptions.