v2.4.2

• UI tweaks
• Custom post processing support
• Audio can be muted/unmuted from main menu (TTS memory usage is 0 while muted)

Full Changelog: v2.4.1...v2.4.2

v2.4.1

• "continue" option no longer selects autosaves (which have to be loaded explicitly now)
• "continue" option shows the name of the story that it will load when selected

Full Changelog: v2.4.0...v2.4.1

v2.4.0 UX update

UX update

• Added text wrapping: words are no longer split in half at the end of each line
• Migrated every input prompt to inquirerpy's new syntax
• Enabled multiline input: enter inserts newline in every case of writing/editing story/context (no more § or ~)
• Replaced every slash command with keyboard shortcuts
• Merged /revert and /edit into a single "undo and edit" feature (key: up arrow)
• Added "redo" feature that allows you to reverse multiple "undo" operations while keeping any edit you made (down arrow)
• Added "edit prompt" feature to edit permanent context after the story has already started
• Added "single sentence generation" feature. Press tab to generate to the end of a sentence instead of all 100 tokens.
• Ctrl+C is now supported in every case: cancels text or audio generation, stops playing audio, clears current input or returns to main menu (depending on the current state) and doesn't terminate the application anymore
• Reworked crash recovery system to create automatic saves at every step of a story, instead of in the case of an exception. This allows reloading after a crash even when python isn't the origin of the termination

Full Changelog: v2.3.2...v2.4.0

v2.3.2

• More languages supported for voice input
• Conversation mode can now generate longer answers
• Voice input model can now be set to run on CPU
• Removed randomized starting prompts
• Minor fixes

Full Changelog: v2.3.1...v2.3.2

v2.3.1

Full Changelog: v2.3.0...v2.3.1

v2.3.0

• Conversation mode: chat with the AI directly
• Mic input: switch to microphone audio input, no need to type!
• Use of EOS token to stop generation early in chat mode (faster chat)
• New /debug command to show some stats about the story, tokens, word cloud
• New /redo command to generate a new output without changing the last input
• Output streaming: words show up in the output as they are generated, and not just when it stops

Full Changelog: 2.2.0...v2.3.0

2.2.0

• Hard limit on vRAM usage (by limiting history) for 24GB AMD GPUs
• EOS token now determined through tokenizer config rather than model config (some models don't declare it the same way)
• Forced the tokenizer to add a leading space if it isn't the default behavior for the model (like Mistral)
• Switched from PyInquirer to InquirerPy, making wrAIter compatible with python 3.10
• Added ROCm requirements for AMD GPUs
• Merged short sentences together during TTS to reduce hallucinations.
• Removed the need for TTS to write temporary audio files to disk, instead everything is played from memory.
• Better crash handling and story recovery after a crash

Full Changelog: v2.1.0...2.2.0

v2.1.0

• help menu fix
• new /save (/s) command to save the story without going to the menu
• quotation marks at the end of generated texts shouldn't be cut as often
• character1 and character2 now correctly go back and forth
• fixed /revert (/r) command when voice was disabled
• stories remember their save name so that it's already filled when saving the same story multiple times
• removed unnecessary first action prompt when creating a new game, it now only prompts for context before running the main loop

Full Changelog: v2.0.2...v2.1.0

v2.0.2 fix

updated requirements

Full Changelog: v2.0.1...v2.0.2

v2.0.1

Introduced the --silent argument to mute all voices.

Full Changelog: v2.0.0...v2.0.1