hume 0.12.1 → 0.13.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (324) hide show
  1. package/.mock/definition/empathic-voice/__package__.yml +760 -711
  2. package/.mock/definition/empathic-voice/chat.yml +29 -23
  3. package/.mock/definition/empathic-voice/chatWebhooks.yml +3 -3
  4. package/.mock/definition/empathic-voice/configs.yml +10 -4
  5. package/.mock/definition/tts/__package__.yml +77 -125
  6. package/.mock/fern.config.json +1 -1
  7. package/api/resources/empathicVoice/resources/chatGroups/client/index.d.ts +1 -0
  8. package/api/resources/empathicVoice/resources/chats/client/index.d.ts +1 -0
  9. package/api/resources/empathicVoice/resources/configs/client/index.d.ts +1 -0
  10. package/api/resources/empathicVoice/resources/configs/client/requests/PostedConfig.d.ts +5 -1
  11. package/api/resources/empathicVoice/resources/prompts/client/index.d.ts +1 -0
  12. package/api/resources/empathicVoice/resources/tools/client/index.d.ts +1 -0
  13. package/api/resources/empathicVoice/types/AssistantEnd.d.ts +2 -2
  14. package/api/resources/empathicVoice/types/AssistantInput.d.ts +2 -2
  15. package/api/resources/empathicVoice/types/AssistantMessage.d.ts +8 -8
  16. package/api/resources/empathicVoice/types/AssistantProsody.d.ts +6 -6
  17. package/api/resources/empathicVoice/types/AudioConfiguration.d.ts +2 -2
  18. package/api/resources/empathicVoice/types/AudioInput.d.ts +6 -6
  19. package/api/resources/empathicVoice/types/AudioOutput.d.ts +4 -4
  20. package/api/resources/empathicVoice/types/BuiltinToolConfig.d.ts +1 -1
  21. package/api/resources/empathicVoice/types/ChatMessage.d.ts +2 -2
  22. package/api/resources/empathicVoice/types/ChatMetadata.d.ts +8 -8
  23. package/api/resources/empathicVoice/types/Context.d.ts +8 -14
  24. package/api/resources/empathicVoice/types/ContextType.d.ts +2 -3
  25. package/api/resources/empathicVoice/types/ContextType.js +1 -2
  26. package/api/resources/empathicVoice/types/LanguageModelType.d.ts +20 -1
  27. package/api/resources/empathicVoice/types/LanguageModelType.js +19 -0
  28. package/api/resources/empathicVoice/types/ModelProviderEnum.d.ts +4 -1
  29. package/api/resources/empathicVoice/types/ModelProviderEnum.js +3 -0
  30. package/api/resources/empathicVoice/types/PauseAssistantMessage.d.ts +2 -2
  31. package/api/resources/empathicVoice/types/ResumeAssistantMessage.d.ts +2 -2
  32. package/api/resources/empathicVoice/types/ReturnChatEvent.d.ts +12 -17
  33. package/api/resources/empathicVoice/types/ReturnChatEventRole.d.ts +0 -4
  34. package/api/resources/empathicVoice/types/ReturnChatEventType.d.ts +22 -18
  35. package/api/resources/empathicVoice/types/ReturnChatEventType.js +9 -4
  36. package/api/resources/empathicVoice/types/ReturnConfig.d.ts +18 -14
  37. package/api/resources/empathicVoice/types/ReturnPrompt.d.ts +2 -2
  38. package/api/resources/empathicVoice/types/SessionSettings.d.ts +29 -29
  39. package/api/resources/empathicVoice/types/Tool.d.ts +6 -6
  40. package/api/resources/empathicVoice/types/ToolCallMessage.d.ts +6 -6
  41. package/api/resources/empathicVoice/types/ToolErrorMessage.d.ts +16 -16
  42. package/api/resources/empathicVoice/types/ToolResponseMessage.d.ts +8 -8
  43. package/api/resources/empathicVoice/types/UserInput.d.ts +2 -2
  44. package/api/resources/empathicVoice/types/UserInterruption.d.ts +4 -4
  45. package/api/resources/empathicVoice/types/UserMessage.d.ts +14 -7
  46. package/api/resources/empathicVoice/types/VoiceProvider.d.ts +2 -1
  47. package/api/resources/empathicVoice/types/VoiceProvider.js +1 -0
  48. package/api/resources/empathicVoice/types/WebSocketError.d.ts +10 -10
  49. package/api/resources/empathicVoice/types/WebhookEventChatEnded.d.ts +8 -8
  50. package/api/resources/empathicVoice/types/WebhookEventChatStarted.d.ts +6 -6
  51. package/api/resources/empathicVoice/types/index.d.ts +16 -16
  52. package/api/resources/empathicVoice/types/index.js +16 -16
  53. package/api/resources/expressionMeasurement/resources/batch/client/Client.d.ts +2 -4
  54. package/api/resources/expressionMeasurement/resources/batch/client/Client.js +1 -1
  55. package/api/resources/expressionMeasurement/resources/batch/client/index.d.ts +1 -0
  56. package/api/resources/tts/client/Client.d.ts +21 -23
  57. package/api/resources/tts/client/Client.js +50 -58
  58. package/api/resources/tts/client/index.d.ts +1 -1
  59. package/api/resources/tts/client/index.js +0 -15
  60. package/api/resources/tts/resources/voices/client/index.d.ts +1 -0
  61. package/api/resources/tts/types/PostedTts.d.ts +8 -8
  62. package/api/resources/tts/types/PostedUtterance.d.ts +6 -6
  63. package/api/resources/tts/types/ReturnGeneration.d.ts +5 -5
  64. package/api/resources/tts/types/ReturnTts.d.ts +1 -1
  65. package/api/resources/tts/types/Snippet.d.ts +6 -6
  66. package/api/resources/tts/types/SnippetAudioChunk.d.ts +0 -16
  67. package/core/fetcher/BinaryResponse.d.ts +17 -0
  68. package/core/fetcher/BinaryResponse.js +14 -0
  69. package/core/fetcher/Fetcher.d.ts +1 -1
  70. package/core/fetcher/ResponseWithBody.d.ts +4 -0
  71. package/core/fetcher/ResponseWithBody.js +6 -0
  72. package/core/fetcher/getFetchFn.js +3 -3
  73. package/core/fetcher/getResponseBody.js +33 -32
  74. package/core/fetcher/index.d.ts +1 -0
  75. package/core/file.d.ts +1 -0
  76. package/core/form-data-utils/FormDataWrapper.d.ts +5 -52
  77. package/core/form-data-utils/FormDataWrapper.js +104 -124
  78. package/core/index.d.ts +1 -0
  79. package/core/index.js +1 -0
  80. package/dist/api/resources/empathicVoice/resources/chatGroups/client/index.d.ts +1 -0
  81. package/dist/api/resources/empathicVoice/resources/chats/client/index.d.ts +1 -0
  82. package/dist/api/resources/empathicVoice/resources/configs/client/index.d.ts +1 -0
  83. package/dist/api/resources/empathicVoice/resources/configs/client/requests/PostedConfig.d.ts +5 -1
  84. package/dist/api/resources/empathicVoice/resources/prompts/client/index.d.ts +1 -0
  85. package/dist/api/resources/empathicVoice/resources/tools/client/index.d.ts +1 -0
  86. package/dist/api/resources/empathicVoice/types/AssistantEnd.d.ts +2 -2
  87. package/dist/api/resources/empathicVoice/types/AssistantInput.d.ts +2 -2
  88. package/dist/api/resources/empathicVoice/types/AssistantMessage.d.ts +8 -8
  89. package/dist/api/resources/empathicVoice/types/AssistantProsody.d.ts +6 -6
  90. package/dist/api/resources/empathicVoice/types/AudioConfiguration.d.ts +2 -2
  91. package/dist/api/resources/empathicVoice/types/AudioInput.d.ts +6 -6
  92. package/dist/api/resources/empathicVoice/types/AudioOutput.d.ts +4 -4
  93. package/dist/api/resources/empathicVoice/types/BuiltinToolConfig.d.ts +1 -1
  94. package/dist/api/resources/empathicVoice/types/ChatMessage.d.ts +2 -2
  95. package/dist/api/resources/empathicVoice/types/ChatMetadata.d.ts +8 -8
  96. package/dist/api/resources/empathicVoice/types/Context.d.ts +8 -14
  97. package/dist/api/resources/empathicVoice/types/ContextType.d.ts +2 -3
  98. package/dist/api/resources/empathicVoice/types/ContextType.js +1 -2
  99. package/dist/api/resources/empathicVoice/types/LanguageModelType.d.ts +20 -1
  100. package/dist/api/resources/empathicVoice/types/LanguageModelType.js +19 -0
  101. package/dist/api/resources/empathicVoice/types/ModelProviderEnum.d.ts +4 -1
  102. package/dist/api/resources/empathicVoice/types/ModelProviderEnum.js +3 -0
  103. package/dist/api/resources/empathicVoice/types/PauseAssistantMessage.d.ts +2 -2
  104. package/dist/api/resources/empathicVoice/types/ResumeAssistantMessage.d.ts +2 -2
  105. package/dist/api/resources/empathicVoice/types/ReturnChatEvent.d.ts +12 -17
  106. package/dist/api/resources/empathicVoice/types/ReturnChatEventRole.d.ts +0 -4
  107. package/dist/api/resources/empathicVoice/types/ReturnChatEventType.d.ts +22 -18
  108. package/dist/api/resources/empathicVoice/types/ReturnChatEventType.js +9 -4
  109. package/dist/api/resources/empathicVoice/types/ReturnConfig.d.ts +18 -14
  110. package/dist/api/resources/empathicVoice/types/ReturnPrompt.d.ts +2 -2
  111. package/dist/api/resources/empathicVoice/types/SessionSettings.d.ts +29 -29
  112. package/dist/api/resources/empathicVoice/types/Tool.d.ts +6 -6
  113. package/dist/api/resources/empathicVoice/types/ToolCallMessage.d.ts +6 -6
  114. package/dist/api/resources/empathicVoice/types/ToolErrorMessage.d.ts +16 -16
  115. package/dist/api/resources/empathicVoice/types/ToolResponseMessage.d.ts +8 -8
  116. package/dist/api/resources/empathicVoice/types/UserInput.d.ts +2 -2
  117. package/dist/api/resources/empathicVoice/types/UserInterruption.d.ts +4 -4
  118. package/dist/api/resources/empathicVoice/types/UserMessage.d.ts +14 -7
  119. package/dist/api/resources/empathicVoice/types/VoiceProvider.d.ts +2 -1
  120. package/dist/api/resources/empathicVoice/types/VoiceProvider.js +1 -0
  121. package/dist/api/resources/empathicVoice/types/WebSocketError.d.ts +10 -10
  122. package/dist/api/resources/empathicVoice/types/WebhookEventChatEnded.d.ts +8 -8
  123. package/dist/api/resources/empathicVoice/types/WebhookEventChatStarted.d.ts +6 -6
  124. package/dist/api/resources/empathicVoice/types/index.d.ts +16 -16
  125. package/dist/api/resources/empathicVoice/types/index.js +16 -16
  126. package/dist/api/resources/expressionMeasurement/resources/batch/client/Client.d.ts +2 -4
  127. package/dist/api/resources/expressionMeasurement/resources/batch/client/Client.js +1 -1
  128. package/dist/api/resources/expressionMeasurement/resources/batch/client/index.d.ts +1 -0
  129. package/dist/api/resources/tts/client/Client.d.ts +21 -23
  130. package/dist/api/resources/tts/client/Client.js +50 -58
  131. package/dist/api/resources/tts/client/index.d.ts +1 -1
  132. package/dist/api/resources/tts/client/index.js +0 -15
  133. package/dist/api/resources/tts/resources/voices/client/index.d.ts +1 -0
  134. package/dist/api/resources/tts/types/PostedTts.d.ts +8 -8
  135. package/dist/api/resources/tts/types/PostedUtterance.d.ts +6 -6
  136. package/dist/api/resources/tts/types/ReturnGeneration.d.ts +5 -5
  137. package/dist/api/resources/tts/types/ReturnTts.d.ts +1 -1
  138. package/dist/api/resources/tts/types/Snippet.d.ts +6 -6
  139. package/dist/api/resources/tts/types/SnippetAudioChunk.d.ts +0 -16
  140. package/dist/core/fetcher/BinaryResponse.d.ts +17 -0
  141. package/dist/core/fetcher/BinaryResponse.js +14 -0
  142. package/dist/core/fetcher/Fetcher.d.ts +1 -1
  143. package/dist/core/fetcher/ResponseWithBody.d.ts +4 -0
  144. package/dist/core/fetcher/ResponseWithBody.js +6 -0
  145. package/dist/core/fetcher/getFetchFn.js +3 -3
  146. package/dist/core/fetcher/getResponseBody.js +33 -32
  147. package/dist/core/fetcher/index.d.ts +1 -0
  148. package/dist/core/file.d.ts +1 -0
  149. package/dist/core/form-data-utils/FormDataWrapper.d.ts +5 -52
  150. package/dist/core/form-data-utils/FormDataWrapper.js +104 -124
  151. package/dist/core/index.d.ts +1 -0
  152. package/dist/core/index.js +1 -0
  153. package/dist/serialization/resources/empathicVoice/types/AssistantEnd.d.ts +1 -1
  154. package/dist/serialization/resources/empathicVoice/types/AssistantEnd.js +1 -1
  155. package/dist/serialization/resources/empathicVoice/types/AssistantInput.d.ts +1 -1
  156. package/dist/serialization/resources/empathicVoice/types/AssistantInput.js +1 -1
  157. package/dist/serialization/resources/empathicVoice/types/AssistantMessage.d.ts +2 -2
  158. package/dist/serialization/resources/empathicVoice/types/AssistantMessage.js +2 -2
  159. package/dist/serialization/resources/empathicVoice/types/AssistantProsody.d.ts +2 -2
  160. package/dist/serialization/resources/empathicVoice/types/AssistantProsody.js +2 -2
  161. package/dist/serialization/resources/empathicVoice/types/AudioConfiguration.d.ts +1 -1
  162. package/dist/serialization/resources/empathicVoice/types/AudioConfiguration.js +1 -1
  163. package/dist/serialization/resources/empathicVoice/types/AudioInput.d.ts +1 -1
  164. package/dist/serialization/resources/empathicVoice/types/AudioInput.js +1 -1
  165. package/dist/serialization/resources/empathicVoice/types/AudioOutput.d.ts +2 -2
  166. package/dist/serialization/resources/empathicVoice/types/AudioOutput.js +2 -2
  167. package/dist/serialization/resources/empathicVoice/types/BuiltinToolConfig.d.ts +1 -1
  168. package/dist/serialization/resources/empathicVoice/types/BuiltinToolConfig.js +1 -1
  169. package/dist/serialization/resources/empathicVoice/types/ChatMessage.d.ts +1 -1
  170. package/dist/serialization/resources/empathicVoice/types/ChatMessage.js +1 -1
  171. package/dist/serialization/resources/empathicVoice/types/ChatMetadata.d.ts +2 -2
  172. package/dist/serialization/resources/empathicVoice/types/ChatMetadata.js +2 -2
  173. package/dist/serialization/resources/empathicVoice/types/Context.d.ts +1 -1
  174. package/dist/serialization/resources/empathicVoice/types/Context.js +1 -1
  175. package/dist/serialization/resources/empathicVoice/types/ContextType.d.ts +1 -1
  176. package/dist/serialization/resources/empathicVoice/types/ContextType.js +1 -1
  177. package/dist/serialization/resources/empathicVoice/types/LanguageModelType.d.ts +1 -1
  178. package/dist/serialization/resources/empathicVoice/types/LanguageModelType.js +19 -0
  179. package/dist/serialization/resources/empathicVoice/types/ModelProviderEnum.d.ts +1 -1
  180. package/dist/serialization/resources/empathicVoice/types/ModelProviderEnum.js +3 -0
  181. package/dist/serialization/resources/empathicVoice/types/PauseAssistantMessage.d.ts +1 -1
  182. package/dist/serialization/resources/empathicVoice/types/PauseAssistantMessage.js +1 -1
  183. package/dist/serialization/resources/empathicVoice/types/ResumeAssistantMessage.d.ts +1 -1
  184. package/dist/serialization/resources/empathicVoice/types/ResumeAssistantMessage.js +1 -1
  185. package/dist/serialization/resources/empathicVoice/types/ReturnChatEventType.d.ts +1 -1
  186. package/dist/serialization/resources/empathicVoice/types/ReturnChatEventType.js +9 -4
  187. package/dist/serialization/resources/empathicVoice/types/ReturnConfig.d.ts +9 -9
  188. package/dist/serialization/resources/empathicVoice/types/ReturnConfig.js +9 -9
  189. package/dist/serialization/resources/empathicVoice/types/ReturnPrompt.d.ts +1 -1
  190. package/dist/serialization/resources/empathicVoice/types/ReturnPrompt.js +1 -1
  191. package/dist/serialization/resources/empathicVoice/types/SessionSettings.d.ts +8 -8
  192. package/dist/serialization/resources/empathicVoice/types/SessionSettings.js +8 -8
  193. package/dist/serialization/resources/empathicVoice/types/Tool.d.ts +3 -3
  194. package/dist/serialization/resources/empathicVoice/types/Tool.js +3 -3
  195. package/dist/serialization/resources/empathicVoice/types/ToolCallMessage.d.ts +3 -3
  196. package/dist/serialization/resources/empathicVoice/types/ToolCallMessage.js +3 -3
  197. package/dist/serialization/resources/empathicVoice/types/ToolErrorMessage.d.ts +6 -6
  198. package/dist/serialization/resources/empathicVoice/types/ToolErrorMessage.js +6 -6
  199. package/dist/serialization/resources/empathicVoice/types/ToolResponseMessage.d.ts +2 -2
  200. package/dist/serialization/resources/empathicVoice/types/ToolResponseMessage.js +2 -2
  201. package/dist/serialization/resources/empathicVoice/types/UserInput.d.ts +1 -1
  202. package/dist/serialization/resources/empathicVoice/types/UserInput.js +1 -1
  203. package/dist/serialization/resources/empathicVoice/types/UserInterruption.d.ts +1 -1
  204. package/dist/serialization/resources/empathicVoice/types/UserInterruption.js +1 -1
  205. package/dist/serialization/resources/empathicVoice/types/UserMessage.d.ts +3 -3
  206. package/dist/serialization/resources/empathicVoice/types/UserMessage.js +3 -3
  207. package/dist/serialization/resources/empathicVoice/types/VoiceProvider.d.ts +1 -1
  208. package/dist/serialization/resources/empathicVoice/types/VoiceProvider.js +1 -1
  209. package/dist/serialization/resources/empathicVoice/types/WebSocketError.d.ts +3 -3
  210. package/dist/serialization/resources/empathicVoice/types/WebSocketError.js +3 -3
  211. package/dist/serialization/resources/empathicVoice/types/WebhookEventChatEnded.d.ts +4 -4
  212. package/dist/serialization/resources/empathicVoice/types/WebhookEventChatEnded.js +4 -4
  213. package/dist/serialization/resources/empathicVoice/types/WebhookEventChatStarted.d.ts +3 -3
  214. package/dist/serialization/resources/empathicVoice/types/WebhookEventChatStarted.js +3 -3
  215. package/dist/serialization/resources/empathicVoice/types/index.d.ts +16 -16
  216. package/dist/serialization/resources/empathicVoice/types/index.js +16 -16
  217. package/dist/serialization/resources/tts/types/PostedTts.d.ts +3 -3
  218. package/dist/serialization/resources/tts/types/PostedTts.js +3 -3
  219. package/dist/serialization/resources/tts/types/PostedUtterance.d.ts +2 -2
  220. package/dist/serialization/resources/tts/types/PostedUtterance.js +2 -2
  221. package/dist/serialization/resources/tts/types/ReturnGeneration.d.ts +3 -3
  222. package/dist/serialization/resources/tts/types/ReturnGeneration.js +3 -3
  223. package/dist/serialization/resources/tts/types/ReturnTts.d.ts +1 -1
  224. package/dist/serialization/resources/tts/types/ReturnTts.js +1 -1
  225. package/dist/serialization/resources/tts/types/Snippet.d.ts +3 -3
  226. package/dist/serialization/resources/tts/types/Snippet.js +3 -3
  227. package/dist/serialization/resources/tts/types/SnippetAudioChunk.d.ts +0 -8
  228. package/dist/serialization/resources/tts/types/SnippetAudioChunk.js +1 -10
  229. package/dist/version.d.ts +1 -1
  230. package/dist/version.js +1 -1
  231. package/jest.browser.config.mjs +10 -0
  232. package/jest.config.mjs +1 -0
  233. package/package.json +6 -7
  234. package/reference.md +25 -27
  235. package/serialization/resources/empathicVoice/types/AssistantEnd.d.ts +1 -1
  236. package/serialization/resources/empathicVoice/types/AssistantEnd.js +1 -1
  237. package/serialization/resources/empathicVoice/types/AssistantInput.d.ts +1 -1
  238. package/serialization/resources/empathicVoice/types/AssistantInput.js +1 -1
  239. package/serialization/resources/empathicVoice/types/AssistantMessage.d.ts +2 -2
  240. package/serialization/resources/empathicVoice/types/AssistantMessage.js +2 -2
  241. package/serialization/resources/empathicVoice/types/AssistantProsody.d.ts +2 -2
  242. package/serialization/resources/empathicVoice/types/AssistantProsody.js +2 -2
  243. package/serialization/resources/empathicVoice/types/AudioConfiguration.d.ts +1 -1
  244. package/serialization/resources/empathicVoice/types/AudioConfiguration.js +1 -1
  245. package/serialization/resources/empathicVoice/types/AudioInput.d.ts +1 -1
  246. package/serialization/resources/empathicVoice/types/AudioInput.js +1 -1
  247. package/serialization/resources/empathicVoice/types/AudioOutput.d.ts +2 -2
  248. package/serialization/resources/empathicVoice/types/AudioOutput.js +2 -2
  249. package/serialization/resources/empathicVoice/types/BuiltinToolConfig.d.ts +1 -1
  250. package/serialization/resources/empathicVoice/types/BuiltinToolConfig.js +1 -1
  251. package/serialization/resources/empathicVoice/types/ChatMessage.d.ts +1 -1
  252. package/serialization/resources/empathicVoice/types/ChatMessage.js +1 -1
  253. package/serialization/resources/empathicVoice/types/ChatMetadata.d.ts +2 -2
  254. package/serialization/resources/empathicVoice/types/ChatMetadata.js +2 -2
  255. package/serialization/resources/empathicVoice/types/Context.d.ts +1 -1
  256. package/serialization/resources/empathicVoice/types/Context.js +1 -1
  257. package/serialization/resources/empathicVoice/types/ContextType.d.ts +1 -1
  258. package/serialization/resources/empathicVoice/types/ContextType.js +1 -1
  259. package/serialization/resources/empathicVoice/types/LanguageModelType.d.ts +1 -1
  260. package/serialization/resources/empathicVoice/types/LanguageModelType.js +19 -0
  261. package/serialization/resources/empathicVoice/types/ModelProviderEnum.d.ts +1 -1
  262. package/serialization/resources/empathicVoice/types/ModelProviderEnum.js +3 -0
  263. package/serialization/resources/empathicVoice/types/PauseAssistantMessage.d.ts +1 -1
  264. package/serialization/resources/empathicVoice/types/PauseAssistantMessage.js +1 -1
  265. package/serialization/resources/empathicVoice/types/ResumeAssistantMessage.d.ts +1 -1
  266. package/serialization/resources/empathicVoice/types/ResumeAssistantMessage.js +1 -1
  267. package/serialization/resources/empathicVoice/types/ReturnChatEventType.d.ts +1 -1
  268. package/serialization/resources/empathicVoice/types/ReturnChatEventType.js +9 -4
  269. package/serialization/resources/empathicVoice/types/ReturnConfig.d.ts +9 -9
  270. package/serialization/resources/empathicVoice/types/ReturnConfig.js +9 -9
  271. package/serialization/resources/empathicVoice/types/ReturnPrompt.d.ts +1 -1
  272. package/serialization/resources/empathicVoice/types/ReturnPrompt.js +1 -1
  273. package/serialization/resources/empathicVoice/types/SessionSettings.d.ts +8 -8
  274. package/serialization/resources/empathicVoice/types/SessionSettings.js +8 -8
  275. package/serialization/resources/empathicVoice/types/Tool.d.ts +3 -3
  276. package/serialization/resources/empathicVoice/types/Tool.js +3 -3
  277. package/serialization/resources/empathicVoice/types/ToolCallMessage.d.ts +3 -3
  278. package/serialization/resources/empathicVoice/types/ToolCallMessage.js +3 -3
  279. package/serialization/resources/empathicVoice/types/ToolErrorMessage.d.ts +6 -6
  280. package/serialization/resources/empathicVoice/types/ToolErrorMessage.js +6 -6
  281. package/serialization/resources/empathicVoice/types/ToolResponseMessage.d.ts +2 -2
  282. package/serialization/resources/empathicVoice/types/ToolResponseMessage.js +2 -2
  283. package/serialization/resources/empathicVoice/types/UserInput.d.ts +1 -1
  284. package/serialization/resources/empathicVoice/types/UserInput.js +1 -1
  285. package/serialization/resources/empathicVoice/types/UserInterruption.d.ts +1 -1
  286. package/serialization/resources/empathicVoice/types/UserInterruption.js +1 -1
  287. package/serialization/resources/empathicVoice/types/UserMessage.d.ts +3 -3
  288. package/serialization/resources/empathicVoice/types/UserMessage.js +3 -3
  289. package/serialization/resources/empathicVoice/types/VoiceProvider.d.ts +1 -1
  290. package/serialization/resources/empathicVoice/types/VoiceProvider.js +1 -1
  291. package/serialization/resources/empathicVoice/types/WebSocketError.d.ts +3 -3
  292. package/serialization/resources/empathicVoice/types/WebSocketError.js +3 -3
  293. package/serialization/resources/empathicVoice/types/WebhookEventChatEnded.d.ts +4 -4
  294. package/serialization/resources/empathicVoice/types/WebhookEventChatEnded.js +4 -4
  295. package/serialization/resources/empathicVoice/types/WebhookEventChatStarted.d.ts +3 -3
  296. package/serialization/resources/empathicVoice/types/WebhookEventChatStarted.js +3 -3
  297. package/serialization/resources/empathicVoice/types/index.d.ts +16 -16
  298. package/serialization/resources/empathicVoice/types/index.js +16 -16
  299. package/serialization/resources/tts/types/PostedTts.d.ts +3 -3
  300. package/serialization/resources/tts/types/PostedTts.js +3 -3
  301. package/serialization/resources/tts/types/PostedUtterance.d.ts +2 -2
  302. package/serialization/resources/tts/types/PostedUtterance.js +2 -2
  303. package/serialization/resources/tts/types/ReturnGeneration.d.ts +3 -3
  304. package/serialization/resources/tts/types/ReturnGeneration.js +3 -3
  305. package/serialization/resources/tts/types/ReturnTts.d.ts +1 -1
  306. package/serialization/resources/tts/types/ReturnTts.js +1 -1
  307. package/serialization/resources/tts/types/Snippet.d.ts +3 -3
  308. package/serialization/resources/tts/types/Snippet.js +3 -3
  309. package/serialization/resources/tts/types/SnippetAudioChunk.d.ts +0 -8
  310. package/serialization/resources/tts/types/SnippetAudioChunk.js +1 -10
  311. package/version.d.ts +1 -1
  312. package/version.js +1 -1
  313. package/api/resources/tts/client/requests/SynthesizeJsonRequest.d.ts +0 -36
  314. package/api/resources/tts/client/requests/SynthesizeJsonRequest.js +0 -5
  315. package/api/resources/tts/client/requests/index.d.ts +0 -1
  316. package/core/form-data-utils/toReadableStream.d.ts +0 -1
  317. package/core/form-data-utils/toReadableStream.js +0 -50
  318. package/dist/api/resources/tts/client/requests/SynthesizeJsonRequest.d.ts +0 -36
  319. package/dist/api/resources/tts/client/requests/SynthesizeJsonRequest.js +0 -5
  320. package/dist/api/resources/tts/client/requests/index.d.ts +0 -1
  321. package/dist/core/form-data-utils/toReadableStream.d.ts +0 -1
  322. package/dist/core/form-data-utils/toReadableStream.js +0 -50
  323. /package/{api/resources/tts/client/requests/index.js → core/file.js} +0 -0
  324. /package/dist/{api/resources/tts/client/requests/index.js → core/file.js} +0 -0
@@ -6,29 +6,15 @@ errors:
6
6
  examples:
7
7
  - value: {}
8
8
  types:
9
- AssistantEnd:
10
- docs: When provided, the output is an assistant end message.
11
- properties:
12
- custom_session_id:
13
- type: optional<string>
14
- docs: >-
15
- Used to manage conversational state, correlate frontend and backend
16
- data, and persist conversations across EVI sessions.
17
- type:
18
- type: literal<"assistant_end">
19
- docs: >-
20
- The type of message sent through the socket; for an Assistant End
21
- message, this must be `assistant_end`.
22
-
23
-
24
- This message indicates the conclusion of the assistant’s response,
25
- signaling that the assistant has finished speaking for the current
26
- conversational turn.
27
- source:
28
- openapi: evi-asyncapi.json
29
9
  AssistantInput:
30
10
  docs: When provided, the input is spoken by EVI.
31
11
  properties:
12
+ type:
13
+ type: literal<"assistant_input">
14
+ docs: >-
15
+ The type of message sent through the socket; must be `assistant_input`
16
+ for our server to correctly identify and process it as an Assistant
17
+ Input message.
32
18
  custom_session_id:
33
19
  type: optional<string>
34
20
  docs: >-
@@ -47,84 +33,16 @@ types:
47
33
  the context of the conversation. The synthesized audio is streamed
48
34
  back to the user as an [Assistant
49
35
  Message](/reference/empathic-voice-interface-evi/chat/chat#receive.AssistantMessage.type).
50
- type:
51
- type: literal<"assistant_input">
52
- docs: >-
53
- The type of message sent through the socket; must be `assistant_input`
54
- for our server to correctly identify and process it as an Assistant
55
- Input message.
56
- source:
57
- openapi: evi-asyncapi.json
58
- AssistantMessage:
59
- docs: When provided, the output is an assistant message.
60
- properties:
61
- custom_session_id:
62
- type: optional<string>
63
- docs: >-
64
- Used to manage conversational state, correlate frontend and backend
65
- data, and persist conversations across EVI sessions.
66
- from_text:
67
- type: boolean
68
- docs: >-
69
- Indicates if this message was inserted into the conversation as text
70
- from an [Assistant Input
71
- message](/reference/empathic-voice-interface-evi/chat/chat#send.AssistantInput.text).
72
- id:
73
- type: optional<string>
74
- docs: >-
75
- ID of the assistant message. Allows the Assistant Message to be
76
- tracked and referenced.
77
- message:
78
- type: ChatMessage
79
- docs: Transcript of the message.
80
- models:
81
- type: Inference
82
- docs: Inference model results.
83
- type:
84
- type: literal<"assistant_message">
85
- docs: >-
86
- The type of message sent through the socket; for an Assistant Message,
87
- this must be `assistant_message`.
88
-
89
-
90
- This message contains both a transcript of the assistant’s response
91
- and the expression measurement predictions of the assistant’s audio
92
- output.
93
- source:
94
- openapi: evi-asyncapi.json
95
- AssistantProsody:
96
- docs: When provided, the output is an Assistant Prosody message.
97
- properties:
98
- custom_session_id:
99
- type: optional<string>
100
- docs: >-
101
- Used to manage conversational state, correlate frontend and backend
102
- data, and persist conversations across EVI sessions.
103
- id:
104
- type: optional<string>
105
- docs: Unique identifier for the segment.
106
- models:
107
- type: Inference
108
- docs: Inference model results.
109
- type:
110
- type: literal<"assistant_prosody">
111
- docs: >-
112
- The type of message sent through the socket; for an Assistant Prosody
113
- message, this must be `assistant_PROSODY`.
114
-
115
-
116
- This message the expression measurement predictions of the assistant's
117
- audio output.
118
36
  source:
119
37
  openapi: evi-asyncapi.json
120
38
  AudioConfiguration:
121
39
  properties:
122
- channels:
123
- type: integer
124
- docs: Number of audio channels.
125
40
  encoding:
126
41
  type: Encoding
127
42
  docs: Encoding format of the audio input, such as `linear16`.
43
+ channels:
44
+ type: integer
45
+ docs: Number of audio channels.
128
46
  sample_rate:
129
47
  type: integer
130
48
  docs: >-
@@ -135,6 +53,17 @@ types:
135
53
  AudioInput:
136
54
  docs: When provided, the input is audio.
137
55
  properties:
56
+ type:
57
+ type: literal<"audio_input">
58
+ docs: >-
59
+ The type of message sent through the socket; must be `audio_input` for
60
+ our server to correctly identify and process it as an Audio Input
61
+ message.
62
+
63
+
64
+ This message is used for sending audio input data to EVI for
65
+ processing and expression measurement. Audio data should be sent as a
66
+ continuous stream, encoded in Base64.
138
67
  custom_session_id:
139
68
  type: optional<string>
140
69
  docs: >-
@@ -157,48 +86,6 @@ types:
157
86
 
158
87
  Hume recommends streaming audio with a buffer window of 20
159
88
  milliseconds (ms), or 100 milliseconds (ms) for web applications.
160
- type:
161
- type: literal<"audio_input">
162
- docs: >-
163
- The type of message sent through the socket; must be `audio_input` for
164
- our server to correctly identify and process it as an Audio Input
165
- message.
166
-
167
-
168
- This message is used for sending audio input data to EVI for
169
- processing and expression measurement. Audio data should be sent as a
170
- continuous stream, encoded in Base64.
171
- source:
172
- openapi: evi-asyncapi.json
173
- AudioOutput:
174
- docs: >-
175
- The type of message sent through the socket; for an Audio Output message,
176
- this must be `audio_output`.
177
- properties:
178
- custom_session_id:
179
- type: optional<string>
180
- docs: >-
181
- Used to manage conversational state, correlate frontend and backend
182
- data, and persist conversations across EVI sessions.
183
- data:
184
- type: string
185
- docs: >-
186
- Base64 encoded audio output. This encoded audio is transmitted to the
187
- client, where it can be decoded and played back as part of the user
188
- interaction.
189
- id:
190
- type: string
191
- docs: >-
192
- ID of the audio output. Allows the Audio Output message to be tracked
193
- and referenced.
194
- index:
195
- type: integer
196
- docs: Index of the chunk of audio relative to the whole audio segment.
197
- type:
198
- type: literal<"audio_output">
199
- docs: >-
200
- The type of message sent through the socket; for an Audio Output
201
- message, this must be `audio_output`.
202
89
  source:
203
90
  openapi: evi-asyncapi.json
204
91
  BuiltInTool:
@@ -212,87 +99,31 @@ types:
212
99
  openapi: evi-asyncapi.json
213
100
  BuiltinToolConfig:
214
101
  properties:
102
+ name:
103
+ type: BuiltInTool
215
104
  fallback_content:
216
105
  type: optional<string>
217
106
  docs: >-
218
107
  Optional text passed to the supplemental LLM if the tool call fails.
219
108
  The LLM then uses this text to generate a response back to the user,
220
109
  ensuring continuity in the conversation.
221
- name:
222
- type: BuiltInTool
223
- source:
224
- openapi: evi-asyncapi.json
225
- ChatMessageToolResult:
226
- discriminated: false
227
- docs: Function call response from client.
228
- union:
229
- - type: ToolResponseMessage
230
- - type: ToolErrorMessage
231
- source:
232
- openapi: evi-asyncapi.json
233
- inline: true
234
- ChatMessage:
235
- properties:
236
- content:
237
- type: optional<string>
238
- docs: Transcript of the message.
239
- role:
240
- type: Role
241
- docs: Role of who is providing the message.
242
- tool_call:
243
- type: optional<ToolCallMessage>
244
- docs: Function call name and arguments.
245
- tool_result:
246
- type: optional<ChatMessageToolResult>
247
- docs: Function call response from client.
248
110
  source:
249
111
  openapi: evi-asyncapi.json
250
- ChatMetadata:
251
- docs: When provided, the output is a chat metadata message.
112
+ Context:
252
113
  properties:
253
- chat_group_id:
254
- type: string
114
+ type:
115
+ type: optional<ContextType>
255
116
  docs: >-
256
- ID of the Chat Group.
257
-
258
-
259
- Used to resume a Chat when passed in the
260
- [resumed_chat_group_id](/reference/empathic-voice-interface-evi/chat/chat#request.query.resumed_chat_group_id)
261
- query parameter of a subsequent connection request. This allows EVI to
262
- continue the conversation from where it left off within the Chat
263
- Group.
117
+ The persistence level of the injected context. Specifies how long the
118
+ injected context will remain active in the session.
264
119
 
265
120
 
266
- Learn more about [supporting chat
267
- resumability](/docs/empathic-voice-interface-evi/faq#does-evi-support-chat-resumability)
268
- from the EVI FAQ.
269
- chat_id:
270
- type: string
271
- docs: >-
272
- ID of the Chat session. Allows the Chat session to be tracked and
273
- referenced.
274
- custom_session_id:
275
- type: optional<string>
276
- docs: >-
277
- Used to manage conversational state, correlate frontend and backend
278
- data, and persist conversations across EVI sessions.
279
- request_id:
280
- type: optional<string>
281
- docs: ID of the initiating request.
282
- type:
283
- type: literal<"chat_metadata">
284
- docs: >-
285
- The type of message sent through the socket; for a Chat Metadata
286
- message, this must be `chat_metadata`.
121
+ - **Temporary**: Context that is only applied to the following
122
+ assistant response.
287
123
 
288
124
 
289
- The Chat Metadata message is the first message you receive after
290
- establishing a connection with EVI and contains important identifiers
291
- for the current Chat session.
292
- source:
293
- openapi: evi-asyncapi.json
294
- Context:
295
- properties:
125
+ - **Persistent**: Context that is applied to all subsequent assistant
126
+ responses for the remainder of the Chat.
296
127
  text:
297
128
  type: string
298
129
  docs: >-
@@ -301,163 +132,29 @@ types:
301
132
  conversation.
302
133
 
303
134
 
304
- This text will be appended to the end of user messages based on the
305
- chosen persistence level. For example, if you want to remind EVI of
306
- its role as a helpful weather assistant, the context you insert will
307
- be appended to the end of user messages as `{Context: You are a
308
- helpful weather assistant}`.
309
- type:
310
- type: optional<ContextType>
311
- docs: >-
312
- The persistence level of the injected context. Specifies how long the
313
- injected context will remain active in the session.
314
-
315
-
316
- There are three possible context types:
317
-
318
-
319
- - **Persistent**: The context is appended to all user messages for the
320
- duration of the session.
321
-
322
-
323
- - **Temporary**: The context is appended only to the next user
324
- message.
325
-
326
- - **Editable**: The original context is updated to reflect the new context.
327
-
328
- If the type is not specified, it will default to `temporary`.
135
+ This text will be appended to the end of
136
+ [user_messages](/reference/speech-to-speech-evi/chat#receive.UserMessage.message.content)
137
+ based on the chosen persistence level. For example, if you want to
138
+ remind EVI of its role as a helpful weather assistant, the context you
139
+ insert will be appended to the end of user messages as `{Context: You
140
+ are a helpful weather assistant}`.
329
141
  source:
330
142
  openapi: evi-asyncapi.json
331
143
  ContextType:
332
144
  enum:
333
- - editable
334
- - persistent
335
145
  - temporary
146
+ - persistent
336
147
  source:
337
148
  openapi: evi-asyncapi.json
338
- EmotionScores:
149
+ Encoding:
150
+ type: literal<"linear16">
151
+ ErrorLevel:
152
+ type: literal<"warn">
153
+ PauseAssistantMessage:
154
+ docs: >-
155
+ Pause responses from EVI. Chat history is still saved and sent after
156
+ resuming.
339
157
  properties:
340
- Admiration: double
341
- Adoration: double
342
- Aesthetic Appreciation: double
343
- Amusement: double
344
- Anger: double
345
- Anxiety: double
346
- Awe: double
347
- Awkwardness: double
348
- Boredom: double
349
- Calmness: double
350
- Concentration: double
351
- Confusion: double
352
- Contemplation: double
353
- Contempt: double
354
- Contentment: double
355
- Craving: double
356
- Desire: double
357
- Determination: double
358
- Disappointment: double
359
- Disgust: double
360
- Distress: double
361
- Doubt: double
362
- Ecstasy: double
363
- Embarrassment: double
364
- Empathic Pain: double
365
- Entrancement: double
366
- Envy: double
367
- Excitement: double
368
- Fear: double
369
- Guilt: double
370
- Horror: double
371
- Interest: double
372
- Joy: double
373
- Love: double
374
- Nostalgia: double
375
- Pain: double
376
- Pride: double
377
- Realization: double
378
- Relief: double
379
- Romance: double
380
- Sadness: double
381
- Satisfaction: double
382
- Shame: double
383
- Surprise (negative): double
384
- Surprise (positive): double
385
- Sympathy: double
386
- Tiredness: double
387
- Triumph: double
388
- source:
389
- openapi: evi-asyncapi.json
390
- Encoding:
391
- type: literal<"linear16">
392
- WebSocketError:
393
- docs: When provided, the output is an error message.
394
- properties:
395
- code:
396
- type: string
397
- docs: Error code. Identifies the type of error encountered.
398
- custom_session_id:
399
- type: optional<string>
400
- docs: >-
401
- Used to manage conversational state, correlate frontend and backend
402
- data, and persist conversations across EVI sessions.
403
- message:
404
- type: string
405
- docs: Detailed description of the error.
406
- request_id:
407
- type: optional<string>
408
- docs: ID of the initiating request.
409
- slug:
410
- type: string
411
- docs: >-
412
- Short, human-readable identifier and description for the error. See a
413
- complete list of error slugs on the [Errors
414
- page](/docs/resources/errors).
415
- type:
416
- type: literal<"error">
417
- docs: >-
418
- The type of message sent through the socket; for a Web Socket Error
419
- message, this must be `error`.
420
-
421
-
422
- This message indicates a disruption in the WebSocket connection, such
423
- as an unexpected disconnection, protocol error, or data transmission
424
- issue.
425
- source:
426
- openapi: evi-asyncapi.json
427
- ErrorLevel:
428
- type: literal<"warn">
429
- Inference:
430
- properties:
431
- prosody:
432
- type: optional<ProsodyInference>
433
- docs: >-
434
- Prosody model inference results.
435
-
436
-
437
- EVI uses the prosody model to measure 48 emotions related to speech
438
- and vocal characteristics within a given expression.
439
- source:
440
- openapi: evi-asyncapi.json
441
- MillisecondInterval:
442
- properties:
443
- begin:
444
- type: integer
445
- docs: Start time of the interval in milliseconds.
446
- end:
447
- type: integer
448
- docs: End time of the interval in milliseconds.
449
- source:
450
- openapi: evi-asyncapi.json
451
- PauseAssistantMessage:
452
- docs: >-
453
- Pause responses from EVI. Chat history is still saved and sent after
454
- resuming.
455
- properties:
456
- custom_session_id:
457
- type: optional<string>
458
- docs: >-
459
- Used to manage conversational state, correlate frontend and backend
460
- data, and persist conversations across EVI sessions.
461
158
  type:
462
159
  type: literal<"pause_assistant_message">
463
160
  docs: >-
@@ -471,24 +168,11 @@ types:
471
168
  message](/reference/empathic-voice-interface-evi/chat/chat#send.ResumeAssistantMessage.type)
472
169
  is sent. When paused, EVI won’t respond, but transcriptions of your
473
170
  audio inputs will still be recorded.
474
- source:
475
- openapi: evi-asyncapi.json
476
- ProsodyInference:
477
- properties:
478
- scores:
479
- type: EmotionScores
171
+ custom_session_id:
172
+ type: optional<string>
480
173
  docs: >-
481
- The confidence scores for 48 emotions within the detected expression
482
- of an audio sample.
483
-
484
-
485
- Scores typically range from 0 to 1, with higher values indicating a
486
- stronger confidence level in the measured attribute.
487
-
488
-
489
- See our guide on [interpreting expression measurement
490
- results](/docs/expression-measurement/faq#how-do-i-interpret-my-results)
491
- to learn more.
174
+ Used to manage conversational state, correlate frontend and backend
175
+ data, and persist conversations across EVI sessions.
492
176
  source:
493
177
  openapi: evi-asyncapi.json
494
178
  ResumeAssistantMessage:
@@ -496,11 +180,6 @@ types:
496
180
  Resume responses from EVI. Chat history sent while paused will now be
497
181
  sent.
498
182
  properties:
499
- custom_session_id:
500
- type: optional<string>
501
- docs: >-
502
- Used to manage conversational state, correlate frontend and backend
503
- data, and persist conversations across EVI sessions.
504
183
  type:
505
184
  type: literal<"resume_assistant_message">
506
185
  docs: >-
@@ -515,15 +194,11 @@ types:
515
194
  then send a `resume_assistant_message`, EVI will respond to the second
516
195
  question and have added the first question to its conversation
517
196
  context.)
518
- source:
519
- openapi: evi-asyncapi.json
520
- Role:
521
- enum:
522
- - assistant
523
- - system
524
- - user
525
- - all
526
- - tool
197
+ custom_session_id:
198
+ type: optional<string>
199
+ docs: >-
200
+ Used to manage conversational state, correlate frontend and backend
201
+ data, and persist conversations across EVI sessions.
527
202
  source:
528
203
  openapi: evi-asyncapi.json
529
204
  SessionSettingsVariablesValue:
@@ -538,48 +213,22 @@ types:
538
213
  SessionSettings:
539
214
  docs: Settings for this chat session.
540
215
  properties:
541
- audio:
542
- type: optional<AudioConfiguration>
543
- docs: >-
544
- Configuration details for the audio input used during the session.
545
- Ensures the audio is being correctly set up for processing.
546
-
547
-
548
- This optional field is only required when the audio input is encoded
549
- in PCM Linear 16 (16-bit, little-endian, signed PCM WAV data). For
550
- detailed instructions on how to configure session settings for PCM
551
- Linear 16 audio, please refer to the [Session Settings
552
- guide](/docs/empathic-voice-interface-evi/configuration/session-settings).
553
- builtin_tools:
554
- type: optional<list<BuiltinToolConfig>>
555
- docs: >-
556
- List of built-in tools to enable for the session.
557
-
558
-
559
- Tools are resources used by EVI to perform various tasks, such as
560
- searching the web or calling external APIs. Built-in tools, like web
561
- search, are natively integrated, while user-defined tools are created
562
- and invoked by the user. To learn more, see our [Tool Use
563
- Guide](/docs/empathic-voice-interface-evi/features/tool-use).
564
-
565
-
566
- Currently, the only built-in tool Hume provides is **Web Search**.
567
- When enabled, Web Search equips EVI with the ability to search the web
568
- for up-to-date information.
569
- context:
570
- type: optional<Context>
216
+ type:
217
+ type: literal<"session_settings">
571
218
  docs: >-
572
- Allows developers to inject additional context into the conversation,
573
- which is appended to the end of user messages for the session.
219
+ The type of message sent through the socket; must be
220
+ `session_settings` for our server to correctly identify and process it
221
+ as a Session Settings message.
574
222
 
575
223
 
576
- When included in a Session Settings message, the provided context can
577
- be used to remind the LLM of its role in every user message, prevent
578
- it from forgetting important details, or add new relevant information
579
- to the conversation.
224
+ Session settings are temporary and apply only to the current Chat
225
+ session. These settings can be adjusted dynamically based on the
226
+ requirements of each session to ensure optimal performance and user
227
+ experience.
580
228
 
581
229
 
582
- Set to `null` to disable context injection.
230
+ For more information, please refer to the [Session Settings
231
+ guide](/docs/empathic-voice-interface-evi/configuration/session-settings).
583
232
  custom_session_id:
584
233
  type: optional<string>
585
234
  docs: >-
@@ -598,17 +247,6 @@ types:
598
247
  language
599
248
  model](/docs/empathic-voice-interface-evi/guides/custom-language-model)
600
249
  with EVI to learn more.
601
- language_model_api_key:
602
- type: optional<string>
603
- docs: >-
604
- Third party API key for the supplemental language model.
605
-
606
-
607
- When provided, EVI will use this key instead of Hume’s API key for the
608
- supplemental LLM. This allows you to bypass rate limits and utilize
609
- your own API key as needed.
610
- metadata:
611
- type: optional<map<string, unknown>>
612
250
  system_prompt:
613
251
  type: optional<string>
614
252
  docs: >-
@@ -631,6 +269,41 @@ types:
631
269
 
632
270
  For help writing a system prompt, see our [Prompting
633
271
  Guide](/docs/empathic-voice-interface-evi/guides/prompting).
272
+ context:
273
+ type: optional<Context>
274
+ docs: >-
275
+ Field for injecting additional context into the conversation, which is
276
+ appended to the end of user messages for the session.
277
+
278
+
279
+ When included in a Session Settings message, the provided context can
280
+ be used to remind the LLM of its role in every user message, prevent
281
+ it from forgetting important details, or add new relevant information
282
+ to the conversation.
283
+
284
+
285
+ Set to `null` to clear injected context.
286
+ audio:
287
+ type: optional<AudioConfiguration>
288
+ docs: >-
289
+ Configuration details for the audio input used during the session.
290
+ Ensures the audio is being correctly set up for processing.
291
+
292
+
293
+ This optional field is only required when the audio input is encoded
294
+ in PCM Linear 16 (16-bit, little-endian, signed PCM WAV data). For
295
+ detailed instructions on how to configure session settings for PCM
296
+ Linear 16 audio, please refer to the [Session Settings
297
+ guide](/docs/empathic-voice-interface-evi/configuration/session-settings).
298
+ language_model_api_key:
299
+ type: optional<string>
300
+ docs: >-
301
+ Third party API key for the supplemental language model.
302
+
303
+
304
+ When provided, EVI will use this key instead of Hume’s API key for the
305
+ supplemental LLM. This allows you to bypass rate limits and utilize
306
+ your own API key as needed.
634
307
  tools:
635
308
  type: optional<list<Tool>>
636
309
  docs: >-
@@ -642,22 +315,24 @@ types:
642
315
  search, are natively integrated, while user-defined tools are created
643
316
  and invoked by the user. To learn more, see our [Tool Use
644
317
  Guide](/docs/empathic-voice-interface-evi/features/tool-use).
645
- type:
646
- type: literal<"session_settings">
318
+ builtin_tools:
319
+ type: optional<list<BuiltinToolConfig>>
647
320
  docs: >-
648
- The type of message sent through the socket; must be
649
- `session_settings` for our server to correctly identify and process it
650
- as a Session Settings message.
321
+ List of built-in tools to enable for the session.
651
322
 
652
323
 
653
- Session settings are temporary and apply only to the current Chat
654
- session. These settings can be adjusted dynamically based on the
655
- requirements of each session to ensure optimal performance and user
656
- experience.
324
+ Tools are resources used by EVI to perform various tasks, such as
325
+ searching the web or calling external APIs. Built-in tools, like web
326
+ search, are natively integrated, while user-defined tools are created
327
+ and invoked by the user. To learn more, see our [Tool Use
328
+ Guide](/docs/empathic-voice-interface-evi/features/tool-use).
657
329
 
658
330
 
659
- For more information, please refer to the [Session Settings
660
- guide](/docs/empathic-voice-interface-evi/configuration/session-settings).
331
+ Currently, the only built-in tool Hume provides is **Web Search**.
332
+ When enabled, Web Search equips EVI with the ability to search the web
333
+ for up-to-date information.
334
+ metadata:
335
+ type: optional<map<string, unknown>>
661
336
  variables:
662
337
  type: optional<map<string, SessionSettingsVariablesValue>>
663
338
  docs: >-
@@ -680,17 +355,9 @@ types:
680
355
  openapi: evi-asyncapi.json
681
356
  Tool:
682
357
  properties:
683
- description:
684
- type: optional<string>
685
- docs: >-
686
- An optional description of what the tool does, used by the
687
- supplemental LLM to choose when and how to call the function.
688
- fallback_content:
689
- type: optional<string>
690
- docs: >-
691
- Optional text passed to the supplemental LLM if the tool call fails.
692
- The LLM then uses this text to generate a response back to the user,
693
- ensuring continuity in the conversation.
358
+ type:
359
+ type: ToolType
360
+ docs: Type of tool. Set to `function` for user-defined tools.
694
361
  name:
695
362
  type: string
696
363
  docs: Name of the user-defined tool to be enabled.
@@ -704,127 +371,87 @@ types:
704
371
  including the expected data type and description for each input field.
705
372
  Structured as a JSON schema, this format ensures the tool receives
706
373
  data in the expected format.
707
- type:
708
- type: ToolType
709
- docs: Type of tool. Set to `function` for user-defined tools.
374
+ description:
375
+ type: optional<string>
376
+ docs: >-
377
+ An optional description of what the tool does, used by the
378
+ supplemental LLM to choose when and how to call the function.
379
+ fallback_content:
380
+ type: optional<string>
381
+ docs: >-
382
+ Optional text passed to the supplemental LLM if the tool call fails.
383
+ The LLM then uses this text to generate a response back to the user,
384
+ ensuring continuity in the conversation.
710
385
  source:
711
386
  openapi: evi-asyncapi.json
712
- ToolCallMessage:
713
- docs: When provided, the output is a tool call.
387
+ ToolErrorMessage:
388
+ docs: When provided, the output is a function call error.
714
389
  properties:
390
+ type:
391
+ type: literal<"tool_error">
392
+ docs: >-
393
+ The type of message sent through the socket; for a Tool Error message,
394
+ this must be `tool_error`.
395
+
396
+
397
+ Upon receiving a [Tool Call
398
+ message](/reference/empathic-voice-interface-evi/chat/chat#receive.ToolCallMessage.type)
399
+ and failing to invoke the function, this message is sent to notify EVI
400
+ of the tool's failure.
715
401
  custom_session_id:
716
402
  type: optional<string>
717
403
  docs: >-
718
404
  Used to manage conversational state, correlate frontend and backend
719
405
  data, and persist conversations across EVI sessions.
720
- name:
721
- type: string
722
- docs: Name of the tool called.
723
- parameters:
724
- type: string
725
- docs: >-
726
- Parameters of the tool.
727
-
728
-
729
- These parameters define the inputs needed for the tool’s execution,
730
- including the expected data type and description for each input field.
731
- Structured as a stringified JSON schema, this format ensures the tool
732
- receives data in the expected format.
733
- response_required:
734
- type: boolean
735
- docs: >-
736
- Indicates whether a response to the tool call is required from the
737
- developer, either in the form of a [Tool Response
738
- message](/reference/empathic-voice-interface-evi/chat/chat#send.ToolResponseMessage.type)
739
- or a [Tool Error
740
- message](/reference/empathic-voice-interface-evi/chat/chat#send.ToolErrorMessage.type).
741
- tool_call_id:
742
- type: string
743
- docs: >-
744
- The unique identifier for a specific tool call instance.
745
-
746
-
747
- This ID is used to track the request and response of a particular tool
748
- invocation, ensuring that the correct response is linked to the
749
- appropriate request.
750
406
  tool_type:
751
407
  type: optional<ToolType>
752
408
  docs: >-
753
409
  Type of tool called. Either `builtin` for natively implemented tools,
754
410
  like web search, or `function` for user-defined tools.
755
- type:
756
- type: literal<"tool_call">
411
+ tool_call_id:
412
+ type: string
757
413
  docs: >-
758
- The type of message sent through the socket; for a Tool Call message,
759
- this must be `tool_call`.
414
+ The unique identifier for a specific tool call instance.
760
415
 
761
416
 
762
- This message indicates that the supplemental LLM has detected a need
763
- to invoke the specified tool.
764
- source:
765
- openapi: evi-asyncapi.json
766
- ToolErrorMessage:
767
- docs: When provided, the output is a function call error.
768
- properties:
769
- code:
770
- type: optional<string>
771
- docs: Error code. Identifies the type of error encountered.
417
+ This ID is used to track the request and response of a particular tool
418
+ invocation, ensuring that the Tool Error message is linked to the
419
+ appropriate tool call request. The specified `tool_call_id` must match
420
+ the one received in the [Tool Call
421
+ message](/reference/empathic-voice-interface-evi/chat/chat#receive.ToolCallMessage.type).
772
422
  content:
773
423
  type: optional<string>
774
424
  docs: >-
775
425
  Optional text passed to the supplemental LLM in place of the tool call
776
426
  result. The LLM then uses this text to generate a response back to the
777
427
  user, ensuring continuity in the conversation if the tool errors.
778
- custom_session_id:
779
- type: optional<string>
780
- docs: >-
781
- Used to manage conversational state, correlate frontend and backend
782
- data, and persist conversations across EVI sessions.
783
428
  error:
784
429
  type: string
785
430
  docs: Error message from the tool call, not exposed to the LLM or user.
431
+ code:
432
+ type: optional<string>
433
+ docs: Error code. Identifies the type of error encountered.
786
434
  level:
787
435
  type: optional<ErrorLevel>
788
436
  docs: >-
789
437
  Indicates the severity of an error; for a Tool Error message, this
790
438
  must be `warn` to signal an unexpected event.
791
- tool_call_id:
792
- type: string
793
- docs: >-
794
- The unique identifier for a specific tool call instance.
795
-
796
-
797
- This ID is used to track the request and response of a particular tool
798
- invocation, ensuring that the Tool Error message is linked to the
799
- appropriate tool call request. The specified `tool_call_id` must match
800
- the one received in the [Tool Call
801
- message](/reference/empathic-voice-interface-evi/chat/chat#receive.ToolCallMessage.type).
802
- tool_type:
803
- type: optional<ToolType>
804
- docs: >-
805
- Type of tool called. Either `builtin` for natively implemented tools,
806
- like web search, or `function` for user-defined tools.
807
- type:
808
- type: literal<"tool_error">
809
- docs: >-
810
- The type of message sent through the socket; for a Tool Error message,
811
- this must be `tool_error`.
812
-
813
-
814
- Upon receiving a [Tool Call
815
- message](/reference/empathic-voice-interface-evi/chat/chat#receive.ToolCallMessage.type)
816
- and failing to invoke the function, this message is sent to notify EVI
817
- of the tool's failure.
818
439
  source:
819
440
  openapi: evi-asyncapi.json
820
441
  ToolResponseMessage:
821
442
  docs: When provided, the output is a function call response.
822
443
  properties:
823
- content:
824
- type: string
444
+ type:
445
+ type: literal<"tool_response">
825
446
  docs: >-
826
- Return value of the tool call. Contains the output generated by the
827
- tool to pass back to EVI.
447
+ The type of message sent through the socket; for a Tool Response
448
+ message, this must be `tool_response`.
449
+
450
+
451
+ Upon receiving a [Tool Call
452
+ message](/reference/empathic-voice-interface-evi/chat/chat#receive.ToolCallMessage.type)
453
+ and successfully invoking the function, this message is sent to convey
454
+ the result of the function call back to EVI.
828
455
  custom_session_id:
829
456
  type: optional<string>
830
457
  docs: >-
@@ -841,6 +468,11 @@ types:
841
468
  appropriate request. The specified `tool_call_id` must match the one
842
469
  received in the [Tool Call
843
470
  message](/reference/empathic-voice-interface-evi/chat/chat#receive.ToolCallMessage.tool_call_id).
471
+ content:
472
+ type: string
473
+ docs: >-
474
+ Return value of the tool call. Contains the output generated by the
475
+ tool to pass back to EVI.
844
476
  tool_name:
845
477
  type: optional<string>
846
478
  docs: >-
@@ -856,17 +488,6 @@ types:
856
488
  docs: >-
857
489
  Type of tool called. Either `builtin` for natively implemented tools,
858
490
  like web search, or `function` for user-defined tools.
859
- type:
860
- type: literal<"tool_response">
861
- docs: >-
862
- The type of message sent through the socket; for a Tool Response
863
- message, this must be `tool_response`.
864
-
865
-
866
- Upon receiving a [Tool Call
867
- message](/reference/empathic-voice-interface-evi/chat/chat#receive.ToolCallMessage.type)
868
- and successfully invoking the function, this message is sent to convey
869
- the result of the function call back to EVI.
870
491
  source:
871
492
  openapi: evi-asyncapi.json
872
493
  ToolType:
@@ -882,44 +503,408 @@ types:
882
503
  and provides a corresponding response.
883
504
 
884
505
 
885
- Expression measurement results are not available for User Input messages,
886
- as the prosody model relies on audio input and cannot process text alone.
887
- properties:
506
+ Expression measurement results are not available for User Input messages,
507
+ as the prosody model relies on audio input and cannot process text alone.
508
+ properties:
509
+ type:
510
+ type: literal<"user_input">
511
+ docs: >-
512
+ The type of message sent through the socket; must be `user_input` for
513
+ our server to correctly identify and process it as a User Input
514
+ message.
515
+ custom_session_id:
516
+ type: optional<string>
517
+ docs: >-
518
+ Used to manage conversational state, correlate frontend and backend
519
+ data, and persist conversations across EVI sessions.
520
+ text:
521
+ type: string
522
+ docs: >-
523
+ User text to insert into the conversation. Text sent through a User
524
+ Input message is treated as the user’s speech to EVI. EVI processes
525
+ this input and provides a corresponding response.
526
+
527
+
528
+ Expression measurement results are not available for User Input
529
+ messages, as the prosody model relies on audio input and cannot
530
+ process text alone.
531
+ source:
532
+ openapi: evi-asyncapi.json
533
+ AssistantEnd:
534
+ docs: When provided, the output is an assistant end message.
535
+ properties:
536
+ type:
537
+ type: literal<"assistant_end">
538
+ docs: >-
539
+ The type of message sent through the socket; for an Assistant End
540
+ message, this must be `assistant_end`.
541
+
542
+
543
+ This message indicates the conclusion of the assistant’s response,
544
+ signaling that the assistant has finished speaking for the current
545
+ conversational turn.
546
+ custom_session_id:
547
+ type: optional<string>
548
+ docs: >-
549
+ Used to manage conversational state, correlate frontend and backend
550
+ data, and persist conversations across EVI sessions.
551
+ source:
552
+ openapi: evi-asyncapi.json
553
+ AssistantMessage:
554
+ docs: When provided, the output is an assistant message.
555
+ properties:
556
+ type:
557
+ type: literal<"assistant_message">
558
+ docs: >-
559
+ The type of message sent through the socket; for an Assistant Message,
560
+ this must be `assistant_message`.
561
+
562
+
563
+ This message contains both a transcript of the assistant’s response
564
+ and the expression measurement predictions of the assistant’s audio
565
+ output.
566
+ custom_session_id:
567
+ type: optional<string>
568
+ docs: >-
569
+ Used to manage conversational state, correlate frontend and backend
570
+ data, and persist conversations across EVI sessions.
571
+ id:
572
+ type: optional<string>
573
+ docs: >-
574
+ ID of the assistant message. Allows the Assistant Message to be
575
+ tracked and referenced.
576
+ message:
577
+ type: ChatMessage
578
+ docs: Transcript of the message.
579
+ models:
580
+ type: Inference
581
+ docs: Inference model results.
582
+ from_text:
583
+ type: boolean
584
+ docs: >-
585
+ Indicates if this message was inserted into the conversation as text
586
+ from an [Assistant Input
587
+ message](/reference/empathic-voice-interface-evi/chat/chat#send.AssistantInput.text).
588
+ source:
589
+ openapi: evi-asyncapi.json
590
+ AssistantProsody:
591
+ docs: When provided, the output is an Assistant Prosody message.
592
+ properties:
593
+ type:
594
+ type: literal<"assistant_prosody">
595
+ docs: >-
596
+ The type of message sent through the socket; for an Assistant Prosody
597
+ message, this must be `assistant_PROSODY`.
598
+
599
+
600
+ This message the expression measurement predictions of the assistant's
601
+ audio output.
602
+ custom_session_id:
603
+ type: optional<string>
604
+ docs: >-
605
+ Used to manage conversational state, correlate frontend and backend
606
+ data, and persist conversations across EVI sessions.
607
+ models:
608
+ type: Inference
609
+ docs: Inference model results.
610
+ id:
611
+ type: optional<string>
612
+ docs: Unique identifier for the segment.
613
+ source:
614
+ openapi: evi-asyncapi.json
615
+ AudioOutput:
616
+ docs: >-
617
+ The type of message sent through the socket; for an Audio Output message,
618
+ this must be `audio_output`.
619
+ properties:
620
+ type:
621
+ type: literal<"audio_output">
622
+ docs: >-
623
+ The type of message sent through the socket; for an Audio Output
624
+ message, this must be `audio_output`.
625
+ custom_session_id:
626
+ type: optional<string>
627
+ docs: >-
628
+ Used to manage conversational state, correlate frontend and backend
629
+ data, and persist conversations across EVI sessions.
630
+ id:
631
+ type: string
632
+ docs: >-
633
+ ID of the audio output. Allows the Audio Output message to be tracked
634
+ and referenced.
635
+ index:
636
+ type: integer
637
+ docs: Index of the chunk of audio relative to the whole audio segment.
638
+ data:
639
+ type: string
640
+ docs: >-
641
+ Base64 encoded audio output. This encoded audio is transmitted to the
642
+ client, where it can be decoded and played back as part of the user
643
+ interaction.
644
+ source:
645
+ openapi: evi-asyncapi.json
646
+ ChatMessageToolResult:
647
+ discriminated: false
648
+ docs: Function call response from client.
649
+ union:
650
+ - type: ToolResponseMessage
651
+ - type: ToolErrorMessage
652
+ source:
653
+ openapi: evi-asyncapi.json
654
+ inline: true
655
+ ChatMessage:
656
+ properties:
657
+ role:
658
+ type: Role
659
+ docs: Role of who is providing the message.
660
+ content:
661
+ type: optional<string>
662
+ docs: Transcript of the message.
663
+ tool_call:
664
+ type: optional<ToolCallMessage>
665
+ docs: Function call name and arguments.
666
+ tool_result:
667
+ type: optional<ChatMessageToolResult>
668
+ docs: Function call response from client.
669
+ source:
670
+ openapi: evi-asyncapi.json
671
+ ChatMetadata:
672
+ docs: When provided, the output is a chat metadata message.
673
+ properties:
674
+ type:
675
+ type: literal<"chat_metadata">
676
+ docs: >-
677
+ The type of message sent through the socket; for a Chat Metadata
678
+ message, this must be `chat_metadata`.
679
+
680
+
681
+ The Chat Metadata message is the first message you receive after
682
+ establishing a connection with EVI and contains important identifiers
683
+ for the current Chat session.
684
+ custom_session_id:
685
+ type: optional<string>
686
+ docs: >-
687
+ Used to manage conversational state, correlate frontend and backend
688
+ data, and persist conversations across EVI sessions.
689
+ chat_group_id:
690
+ type: string
691
+ docs: >-
692
+ ID of the Chat Group.
693
+
694
+
695
+ Used to resume a Chat when passed in the
696
+ [resumed_chat_group_id](/reference/empathic-voice-interface-evi/chat/chat#request.query.resumed_chat_group_id)
697
+ query parameter of a subsequent connection request. This allows EVI to
698
+ continue the conversation from where it left off within the Chat
699
+ Group.
700
+
701
+
702
+ Learn more about [supporting chat
703
+ resumability](/docs/empathic-voice-interface-evi/faq#does-evi-support-chat-resumability)
704
+ from the EVI FAQ.
705
+ chat_id:
706
+ type: string
707
+ docs: >-
708
+ ID of the Chat session. Allows the Chat session to be tracked and
709
+ referenced.
710
+ request_id:
711
+ type: optional<string>
712
+ docs: ID of the initiating request.
713
+ source:
714
+ openapi: evi-asyncapi.json
715
+ EmotionScores:
716
+ properties:
717
+ Admiration: double
718
+ Adoration: double
719
+ Aesthetic Appreciation: double
720
+ Amusement: double
721
+ Anger: double
722
+ Anxiety: double
723
+ Awe: double
724
+ Awkwardness: double
725
+ Boredom: double
726
+ Calmness: double
727
+ Concentration: double
728
+ Confusion: double
729
+ Contemplation: double
730
+ Contempt: double
731
+ Contentment: double
732
+ Craving: double
733
+ Desire: double
734
+ Determination: double
735
+ Disappointment: double
736
+ Disgust: double
737
+ Distress: double
738
+ Doubt: double
739
+ Ecstasy: double
740
+ Embarrassment: double
741
+ Empathic Pain: double
742
+ Entrancement: double
743
+ Envy: double
744
+ Excitement: double
745
+ Fear: double
746
+ Guilt: double
747
+ Horror: double
748
+ Interest: double
749
+ Joy: double
750
+ Love: double
751
+ Nostalgia: double
752
+ Pain: double
753
+ Pride: double
754
+ Realization: double
755
+ Relief: double
756
+ Romance: double
757
+ Sadness: double
758
+ Satisfaction: double
759
+ Shame: double
760
+ Surprise (negative): double
761
+ Surprise (positive): double
762
+ Sympathy: double
763
+ Tiredness: double
764
+ Triumph: double
765
+ source:
766
+ openapi: evi-asyncapi.json
767
+ WebSocketError:
768
+ docs: When provided, the output is an error message.
769
+ properties:
770
+ type:
771
+ type: literal<"error">
772
+ docs: >-
773
+ The type of message sent through the socket; for a Web Socket Error
774
+ message, this must be `error`.
775
+
776
+
777
+ This message indicates a disruption in the WebSocket connection, such
778
+ as an unexpected disconnection, protocol error, or data transmission
779
+ issue.
780
+ custom_session_id:
781
+ type: optional<string>
782
+ docs: >-
783
+ Used to manage conversational state, correlate frontend and backend
784
+ data, and persist conversations across EVI sessions.
785
+ code:
786
+ type: string
787
+ docs: Error code. Identifies the type of error encountered.
788
+ slug:
789
+ type: string
790
+ docs: >-
791
+ Short, human-readable identifier and description for the error. See a
792
+ complete list of error slugs on the [Errors
793
+ page](/docs/resources/errors).
794
+ message:
795
+ type: string
796
+ docs: Detailed description of the error.
797
+ request_id:
798
+ type: optional<string>
799
+ docs: ID of the initiating request.
800
+ source:
801
+ openapi: evi-asyncapi.json
802
+ Inference:
803
+ properties:
804
+ prosody:
805
+ type: optional<ProsodyInference>
806
+ docs: >-
807
+ Prosody model inference results.
808
+
809
+
810
+ EVI uses the prosody model to measure 48 emotions related to speech
811
+ and vocal characteristics within a given expression.
812
+ source:
813
+ openapi: evi-asyncapi.json
814
+ MillisecondInterval:
815
+ properties:
816
+ begin:
817
+ type: integer
818
+ docs: Start time of the interval in milliseconds.
819
+ end:
820
+ type: integer
821
+ docs: End time of the interval in milliseconds.
822
+ source:
823
+ openapi: evi-asyncapi.json
824
+ ProsodyInference:
825
+ properties:
826
+ scores:
827
+ type: EmotionScores
828
+ docs: >-
829
+ The confidence scores for 48 emotions within the detected expression
830
+ of an audio sample.
831
+
832
+
833
+ Scores typically range from 0 to 1, with higher values indicating a
834
+ stronger confidence level in the measured attribute.
835
+
836
+
837
+ See our guide on [interpreting expression measurement
838
+ results](/docs/expression-measurement/faq#how-do-i-interpret-my-results)
839
+ to learn more.
840
+ source:
841
+ openapi: evi-asyncapi.json
842
+ Role:
843
+ enum:
844
+ - assistant
845
+ - system
846
+ - user
847
+ - all
848
+ - tool
849
+ source:
850
+ openapi: evi-asyncapi.json
851
+ ToolCallMessage:
852
+ docs: When provided, the output is a tool call.
853
+ properties:
854
+ name:
855
+ type: string
856
+ docs: Name of the tool called.
857
+ parameters:
858
+ type: string
859
+ docs: >-
860
+ Parameters of the tool.
861
+
862
+
863
+ These parameters define the inputs needed for the tool’s execution,
864
+ including the expected data type and description for each input field.
865
+ Structured as a stringified JSON schema, this format ensures the tool
866
+ receives data in the expected format.
867
+ tool_call_id:
868
+ type: string
869
+ docs: >-
870
+ The unique identifier for a specific tool call instance.
871
+
872
+
873
+ This ID is used to track the request and response of a particular tool
874
+ invocation, ensuring that the correct response is linked to the
875
+ appropriate request.
876
+ type:
877
+ type: literal<"tool_call">
878
+ docs: >-
879
+ The type of message sent through the socket; for a Tool Call message,
880
+ this must be `tool_call`.
881
+
882
+
883
+ This message indicates that the supplemental LLM has detected a need
884
+ to invoke the specified tool.
888
885
  custom_session_id:
889
886
  type: optional<string>
890
887
  docs: >-
891
888
  Used to manage conversational state, correlate frontend and backend
892
889
  data, and persist conversations across EVI sessions.
893
- text:
894
- type: string
890
+ tool_type:
891
+ type: optional<ToolType>
895
892
  docs: >-
896
- User text to insert into the conversation. Text sent through a User
897
- Input message is treated as the user’s speech to EVI. EVI processes
898
- this input and provides a corresponding response.
899
-
900
-
901
- Expression measurement results are not available for User Input
902
- messages, as the prosody model relies on audio input and cannot
903
- process text alone.
904
- type:
905
- type: literal<"user_input">
893
+ Type of tool called. Either `builtin` for natively implemented tools,
894
+ like web search, or `function` for user-defined tools.
895
+ response_required:
896
+ type: boolean
906
897
  docs: >-
907
- The type of message sent through the socket; must be `user_input` for
908
- our server to correctly identify and process it as a User Input
909
- message.
898
+ Indicates whether a response to the tool call is required from the
899
+ developer, either in the form of a [Tool Response
900
+ message](/reference/empathic-voice-interface-evi/chat/chat#send.ToolResponseMessage.type)
901
+ or a [Tool Error
902
+ message](/reference/empathic-voice-interface-evi/chat/chat#send.ToolErrorMessage.type).
910
903
  source:
911
904
  openapi: evi-asyncapi.json
912
905
  UserInterruption:
913
906
  docs: When provided, the output is an interruption.
914
907
  properties:
915
- custom_session_id:
916
- type: optional<string>
917
- docs: >-
918
- Used to manage conversational state, correlate frontend and backend
919
- data, and persist conversations across EVI sessions.
920
- time:
921
- type: integer
922
- docs: Unix timestamp of the detected user interruption.
923
908
  type:
924
909
  type: literal<"user_interruption">
925
910
  docs: >-
@@ -932,40 +917,39 @@ types:
932
917
  message to signal the interruption event. This message allows the
933
918
  system to stop the current audio playback, clear the audio queue, and
934
919
  prepare to handle new user input.
920
+ custom_session_id:
921
+ type: optional<string>
922
+ docs: >-
923
+ Used to manage conversational state, correlate frontend and backend
924
+ data, and persist conversations across EVI sessions.
925
+ time:
926
+ type: integer
927
+ docs: Unix timestamp of the detected user interruption.
935
928
  source:
936
929
  openapi: evi-asyncapi.json
937
930
  UserMessage:
938
931
  docs: When provided, the output is a user message.
939
932
  properties:
933
+ type:
934
+ type: literal<"user_message">
935
+ docs: >-
936
+ The type of message sent through the socket; for a User Message, this
937
+ must be `user_message`.
938
+
939
+
940
+ This message contains both a transcript of the user’s input and the
941
+ expression measurement predictions if the input was sent as an [Audio
942
+ Input
943
+ message](/reference/empathic-voice-interface-evi/chat/chat#send.AudioInput.type).
944
+ Expression measurement predictions are not provided for a [User Input
945
+ message](/reference/empathic-voice-interface-evi/chat/chat#send.UserInput.type),
946
+ as the prosody model relies on audio input and cannot process text
947
+ alone.
940
948
  custom_session_id:
941
949
  type: optional<string>
942
950
  docs: >-
943
951
  Used to manage conversational state, correlate frontend and backend
944
952
  data, and persist conversations across EVI sessions.
945
- from_text:
946
- type: boolean
947
- docs: >-
948
- Indicates if this message was inserted into the conversation as text
949
- from a [User
950
- Input](/reference/empathic-voice-interface-evi/chat/chat#send.UserInput.text)
951
- message.
952
- interim:
953
- type: boolean
954
- docs: >-
955
- Indicates if this message contains an immediate and unfinalized
956
- transcript of the user’s audio input. If it does, words may be
957
- repeated across successive `UserMessage` messages as our transcription
958
- model becomes more confident about what was said with additional
959
- context. Interim messages are useful to detect if the user is
960
- interrupting during audio playback on the client. Even without a
961
- finalized transcription, along with
962
- [UserInterrupt](/reference/empathic-voice-interface-evi/chat/chat#receive.UserInterruption.type)
963
- messages, interim `UserMessages` are useful for detecting if the user
964
- is interrupting during audio playback on the client, signaling to stop
965
- playback in your application. Interim `UserMessages` will only be
966
- received if the
967
- [verbose_transcription](/reference/empathic-voice-interface-evi/chat/chat#request.query.verbose_transcription)
968
- query parameter is set to `true` in the handshake request.
969
953
  message:
970
954
  type: ChatMessage
971
955
  docs: Transcript of the message.
@@ -975,21 +959,30 @@ types:
975
959
  time:
976
960
  type: MillisecondInterval
977
961
  docs: Start and End time of user message.
978
- type:
979
- type: literal<"user_message">
962
+ from_text:
963
+ type: boolean
980
964
  docs: >-
981
- The type of message sent through the socket; for a User Message, this
982
- must be `user_message`.
965
+ Indicates if this message was inserted into the conversation as text
966
+ from a [User
967
+ Input](/reference/empathic-voice-interface-evi/chat/chat#send.UserInput.text)
968
+ message.
969
+ interim:
970
+ type: boolean
971
+ docs: >-
972
+ Indicates whether this `UserMessage` contains an interim (unfinalized)
973
+ transcript.
983
974
 
984
975
 
985
- This message contains both a transcript of the user’s input and the
986
- expression measurement predictions if the input was sent as an [Audio
987
- Input
988
- message](/reference/empathic-voice-interface-evi/chat/chat#send.AudioInput.type).
989
- Expression measurement predictions are not provided for a [User Input
990
- message](/reference/empathic-voice-interface-evi/chat/chat#send.UserInput.type),
991
- as the prosody model relies on audio input and cannot process text
992
- alone.
976
+ - `true`: the transcript is provisional; words may be repeated or
977
+ refined in subsequent `UserMessage` responses as additional audio is
978
+ processed.
979
+
980
+ - `false`: the transcript is final and complete.
981
+
982
+
983
+ Interim transcripts are only sent when the
984
+ [`verbose_transcription`](/reference/empathic-voice-interface-evi/chat/chat#request.query.verbose_transcription)
985
+ query parameter is set to `true` in the initial handshake.
993
986
  source:
994
987
  openapi: evi-asyncapi.json
995
988
  JsonMessage:
@@ -1029,12 +1022,22 @@ types:
1029
1022
  name: Claude3Sonnet20240229
1030
1023
  - value: claude-3-haiku-20240307
1031
1024
  name: Claude3Haiku20240307
1025
+ - value: claude-sonnet-4-20250514
1026
+ name: ClaudeSonnet420250514
1032
1027
  - value: us.anthropic.claude-3-5-haiku-20241022-v1:0
1033
1028
  name: UsAnthropicClaude35Haiku20241022V10
1034
1029
  - value: us.anthropic.claude-3-5-sonnet-20240620-v1:0
1035
1030
  name: UsAnthropicClaude35Sonnet20240620V10
1036
1031
  - value: us.anthropic.claude-3-haiku-20240307-v1:0
1037
1032
  name: UsAnthropicClaude3Haiku20240307V10
1033
+ - value: gpt-oss-120b
1034
+ name: GptOss120B
1035
+ - value: qwen-3-235b-a22b
1036
+ name: Qwen3235BA22B
1037
+ - value: qwen-3-235b-a22b-instruct-2507
1038
+ name: Qwen3235BA22BInstruct2507
1039
+ - value: qwen-3-235b-a22b-thinking-2507
1040
+ name: Qwen3235BA22BThinking2507
1038
1041
  - value: gemini-1.5-pro
1039
1042
  name: Gemini15Pro
1040
1043
  - value: gemini-1.5-flash
@@ -1045,6 +1048,10 @@ types:
1045
1048
  name: Gemini15Flash002
1046
1049
  - value: gemini-2.0-flash
1047
1050
  name: Gemini20Flash
1051
+ - value: gemini-2.5-flash
1052
+ name: Gemini25Flash
1053
+ - value: gemini-2.5-flash-preview-04-17
1054
+ name: Gemini25FlashPreview0417
1048
1055
  - value: gpt-4-turbo
1049
1056
  name: Gpt4Turbo
1050
1057
  - value: gpt-4-turbo-preview
@@ -1057,6 +1064,14 @@ types:
1057
1064
  name: Gpt4O
1058
1065
  - value: gpt-4o-mini
1059
1066
  name: Gpt4OMini
1067
+ - value: gpt-4.1
1068
+ name: Gpt41
1069
+ - value: gpt-5
1070
+ name: Gpt5
1071
+ - value: gpt-5-mini
1072
+ name: Gpt5Mini
1073
+ - value: gpt-5-nano
1074
+ name: Gpt5Nano
1060
1075
  - value: gemma-7b-it
1061
1076
  name: Gemma7BIt
1062
1077
  - value: llama3-8b-8192
@@ -1069,6 +1084,8 @@ types:
1069
1084
  name: Llama3370BVersatile
1070
1085
  - value: llama-3.1-8b-instant
1071
1086
  name: Llama318BInstant
1087
+ - value: moonshotai/kimi-k2-instruct
1088
+ name: MoonshotaiKimiK2Instruct
1072
1089
  - value: accounts/fireworks/models/mixtral-8x7b-instruct
1073
1090
  name: AccountsFireworksModelsMixtral8X7BInstruct
1074
1091
  - value: accounts/fireworks/models/llama-v3p1-405b-instruct
@@ -1077,9 +1094,21 @@ types:
1077
1094
  name: AccountsFireworksModelsLlamaV3P170BInstruct
1078
1095
  - value: accounts/fireworks/models/llama-v3p1-8b-instruct
1079
1096
  name: AccountsFireworksModelsLlamaV3P18BInstruct
1097
+ - sonar
1098
+ - value: sonar-pro
1099
+ name: SonarPro
1100
+ - sambanova
1101
+ - value: DeepSeek-R1-Distill-Llama-70B
1102
+ name: DeepSeekR1DistillLlama70B
1103
+ - value: Llama-4-Maverick-17B-128E-Instruct
1104
+ name: Llama4Maverick17B128EInstruct
1105
+ - value: Qwen3-32B
1106
+ name: Qwen332B
1080
1107
  - ellm
1081
1108
  - value: custom-language-model
1082
1109
  name: CustomLanguageModel
1110
+ - value: hume-evi-3-web-search
1111
+ name: HumeEvi3WebSearch
1083
1112
  source:
1084
1113
  openapi: evi-openapi.json
1085
1114
  ModelProviderEnum:
@@ -1092,6 +1121,9 @@ types:
1092
1121
  - GOOGLE
1093
1122
  - HUME_AI
1094
1123
  - AMAZON_BEDROCK
1124
+ - PERPLEXITY
1125
+ - SAMBANOVA
1126
+ - CEREBRAS
1095
1127
  source:
1096
1128
  openapi: evi-openapi.json
1097
1129
  ValidationErrorLocItem:
@@ -1110,29 +1142,20 @@ types:
1110
1142
  type: string
1111
1143
  source:
1112
1144
  openapi: evi-openapi.json
1113
- WebhookEventBase:
1114
- docs: Represents the fields common to all webhook events.
1115
- properties:
1116
- chat_group_id:
1117
- type: string
1118
- docs: Unique ID of the **Chat Group** associated with the **Chat** session.
1119
- chat_id:
1120
- type: string
1121
- docs: Unique ID of the **Chat** session.
1122
- config_id:
1123
- type: optional<string>
1124
- docs: Unique ID of the EVI **Config** used for the session.
1125
- source:
1126
- openapi: evi-openapi.json
1127
- WebhookEvent:
1128
- discriminated: false
1129
- union:
1130
- - WebhookEventChatStarted
1131
- - WebhookEventChatEnded
1132
- source:
1133
- openapi: evi-openapi.json
1134
1145
  WebhookEventChatEnded:
1135
1146
  properties:
1147
+ event_name:
1148
+ type: optional<literal<"chat_ended">>
1149
+ docs: Always `chat_ended`.
1150
+ end_time:
1151
+ type: integer
1152
+ docs: Unix timestamp (in milliseconds) indicating when the session ended.
1153
+ duration_seconds:
1154
+ type: integer
1155
+ docs: Total duration of the session in seconds.
1156
+ end_reason:
1157
+ type: WebhookEventChatStatus
1158
+ docs: Reason for the session's termination.
1136
1159
  caller_number:
1137
1160
  type: optional<string>
1138
1161
  docs: >-
@@ -1147,18 +1170,6 @@ types:
1147
1170
  language
1148
1171
  model](/docs/empathic-voice-interface-evi/custom-language-model) in
1149
1172
  the EVI Config.
1150
- duration_seconds:
1151
- type: integer
1152
- docs: Total duration of the session in seconds.
1153
- end_reason:
1154
- type: WebhookEventChatStatus
1155
- docs: Reason for the session's termination.
1156
- end_time:
1157
- type: integer
1158
- docs: Unix timestamp (in milliseconds) indicating when the session ended.
1159
- event_name:
1160
- type: optional<literal<"chat_ended">>
1161
- docs: Always `chat_ended`.
1162
1173
  extends:
1163
1174
  - WebhookEventBase
1164
1175
  source:
@@ -1171,6 +1182,18 @@ types:
1171
1182
  openapi: evi-openapi.json
1172
1183
  WebhookEventChatStarted:
1173
1184
  properties:
1185
+ event_name:
1186
+ type: optional<literal<"chat_started">>
1187
+ docs: Always `chat_started`.
1188
+ start_time:
1189
+ type: integer
1190
+ docs: Unix timestamp (in milliseconds) indicating when the session started.
1191
+ chat_start_type:
1192
+ type: WebhookEventChatStartType
1193
+ docs: >-
1194
+ Indicates whether the chat is the first in a new Chat Group
1195
+ (`new_chat_group`) or the continuation of an existing chat group
1196
+ (`resumed_chat_group`).
1174
1197
  caller_number:
1175
1198
  type: optional<string>
1176
1199
  docs: >-
@@ -1178,12 +1201,6 @@ types:
1178
1201
  This field is included only if the Chat was created via the [Twilio
1179
1202
  phone calling](/docs/empathic-voice-interface-evi/phone-calling)
1180
1203
  integration.
1181
- chat_start_type:
1182
- type: WebhookEventChatStartType
1183
- docs: >-
1184
- Indicates whether the chat is the first in a new Chat Group
1185
- (`new_chat_group`) or the continuation of an existing chat group
1186
- (`resumed_chat_group`).
1187
1204
  custom_session_id:
1188
1205
  type: optional<string>
1189
1206
  docs: >-
@@ -1191,12 +1208,6 @@ types:
1191
1208
  language
1192
1209
  model](/docs/empathic-voice-interface-evi/custom-language-model) in
1193
1210
  the EVI Config.
1194
- event_name:
1195
- type: optional<literal<"chat_started">>
1196
- docs: Always `chat_started`.
1197
- start_time:
1198
- type: integer
1199
- docs: Unix timestamp (in milliseconds) indicating when the session started.
1200
1211
  extends:
1201
1212
  - WebhookEventBase
1202
1213
  source:
@@ -1212,6 +1223,27 @@ types:
1212
1223
  - ERROR
1213
1224
  source:
1214
1225
  openapi: evi-openapi.json
1226
+ WebhookEvent:
1227
+ discriminated: false
1228
+ union:
1229
+ - WebhookEventChatStarted
1230
+ - WebhookEventChatEnded
1231
+ source:
1232
+ openapi: evi-openapi.json
1233
+ WebhookEventBase:
1234
+ docs: Represents the fields common to all webhook events.
1235
+ properties:
1236
+ chat_group_id:
1237
+ type: string
1238
+ docs: Unique ID of the **Chat Group** associated with the **Chat** session.
1239
+ chat_id:
1240
+ type: string
1241
+ docs: Unique ID of the **Chat** session.
1242
+ config_id:
1243
+ type: optional<string>
1244
+ docs: Unique ID of the EVI **Config** used for the session.
1245
+ source:
1246
+ openapi: evi-openapi.json
1215
1247
  ErrorResponse:
1216
1248
  properties:
1217
1249
  error: optional<string>
@@ -1405,6 +1437,9 @@ types:
1405
1437
  Version numbers are integer values representing different iterations
1406
1438
  of the Prompt. Each update to the Prompt increments its version
1407
1439
  number.
1440
+ version_description:
1441
+ type: optional<string>
1442
+ docs: An optional description of the Prompt version.
1408
1443
  version_type:
1409
1444
  type: ReturnPromptVersionType
1410
1445
  docs: >-
@@ -1420,9 +1455,6 @@ types:
1420
1455
  docs: >-
1421
1456
  Time at which the Prompt was last modified. Measured in seconds since
1422
1457
  the Unix epoch.
1423
- version_description:
1424
- type: optional<string>
1425
- docs: An optional description of the Prompt version.
1426
1458
  source:
1427
1459
  openapi: evi-openapi.json
1428
1460
  ReturnPagedConfigs:
@@ -1478,15 +1510,39 @@ types:
1478
1510
  Version numbers are integer values representing different iterations
1479
1511
  of the Config. Each update to the Config increments its version
1480
1512
  number.
1513
+ tools:
1514
+ type: optional<list<optional<ReturnUserDefinedTool>>>
1515
+ docs: List of user-defined tools associated with this Config.
1516
+ version_description:
1517
+ type: optional<string>
1518
+ docs: An optional description of the Config version.
1519
+ language_model:
1520
+ type: optional<ReturnLanguageModel>
1521
+ docs: >-
1522
+ The supplemental language model associated with this Config.
1523
+
1524
+
1525
+ This model is used to generate longer, more detailed responses from
1526
+ EVI. Choosing an appropriate supplemental language model for your use
1527
+ case is crucial for generating fast, high-quality responses from EVI.
1528
+ builtin_tools:
1529
+ type: optional<list<optional<ReturnBuiltinTool>>>
1530
+ docs: List of built-in tools associated with this Config.
1481
1531
  evi_version:
1482
1532
  type: optional<string>
1483
1533
  docs: >-
1484
- Specifies the EVI version to use. Use `"1"` for version 1, or `"2"`
1485
- for the latest enhanced version. For a detailed comparison of the two
1486
- versions, refer to our
1487
- [guide](/docs/speech-to-speech-evi/configuration/evi-version).
1534
+ Specifies the EVI version to use. See our [EVI Version
1535
+ Guide](/docs/speech-to-speech-evi/configuration/evi-version) for
1536
+ differences between versions.
1537
+
1538
+
1539
+ **We're officially sunsetting EVI versions 1 and 2 on August 30,
1540
+ 2025**. To keep things running smoothly, be sure to [migrate to EVI
1541
+ 3](/docs/speech-to-speech-evi/configuration/evi-version#migrating-to-evi-3)
1542
+ before then.
1488
1543
  timeouts: optional<ReturnTimeoutSpecs>
1489
1544
  nudges: optional<ReturnNudgeSpec>
1545
+ event_messages: optional<ReturnEventMessageSpecs>
1490
1546
  ellm_model:
1491
1547
  type: optional<ReturnEllmModel>
1492
1548
  docs: >-
@@ -1499,9 +1555,6 @@ types:
1499
1555
  text-to-speech (TTS) prosody.
1500
1556
  voice: optional<unknown>
1501
1557
  prompt: optional<ReturnPrompt>
1502
- tools:
1503
- type: optional<list<optional<ReturnUserDefinedTool>>>
1504
- docs: List of user-defined tools associated with this Config.
1505
1558
  webhooks:
1506
1559
  type: optional<list<optional<ReturnWebhookSpec>>>
1507
1560
  docs: Map of webhooks associated with this config.
@@ -1515,22 +1568,6 @@ types:
1515
1568
  docs: >-
1516
1569
  Time at which the Config was last modified. Measured in seconds since
1517
1570
  the Unix epoch.
1518
- language_model:
1519
- type: optional<ReturnLanguageModel>
1520
- docs: >-
1521
- The supplemental language model associated with this Config.
1522
-
1523
-
1524
- This model is used to generate longer, more detailed responses from
1525
- EVI. Choosing an appropriate supplemental language model for your use
1526
- case is crucial for generating fast, high-quality responses from EVI.
1527
- builtin_tools:
1528
- type: optional<list<optional<ReturnBuiltinTool>>>
1529
- docs: List of built-in tools associated with this Config.
1530
- event_messages: optional<ReturnEventMessageSpecs>
1531
- version_description:
1532
- type: optional<string>
1533
- docs: An optional description of the Config version.
1534
1571
  source:
1535
1572
  openapi: evi-openapi.json
1536
1573
  ReturnPagedChatsPaginationDirection:
@@ -2632,50 +2669,58 @@ types:
2632
2669
  The role of the entity which generated the Chat Event. There are four
2633
2670
  possible values:
2634
2671
 
2635
-
2636
2672
  - `USER`: The user, capable of sending user messages and interruptions.
2637
2673
 
2638
-
2639
2674
  - `AGENT`: The assistant, capable of sending agent messages.
2640
2675
 
2641
-
2642
2676
  - `SYSTEM`: The backend server, capable of transmitting errors.
2643
2677
 
2644
-
2645
2678
  - `TOOL`: The function calling mechanism.
2646
2679
  inline: true
2647
2680
  source:
2648
2681
  openapi: evi-openapi.json
2649
2682
  ReturnChatEventType:
2650
2683
  enum:
2651
- - SYSTEM_PROMPT
2652
- - USER_MESSAGE
2653
- - USER_INTERRUPTION
2654
- - AGENT_MESSAGE
2655
2684
  - FUNCTION_CALL
2656
2685
  - FUNCTION_CALL_RESPONSE
2686
+ - CHAT_END_MESSAGE
2687
+ - AGENT_MESSAGE
2688
+ - SYSTEM_PROMPT
2689
+ - USER_RECORDING_START_MESSAGE
2690
+ - RESUME_ONSET
2691
+ - USER_INTERRUPTION
2692
+ - CHAT_START_MESSAGE
2693
+ - PAUSE_ONSET
2694
+ - USER_MESSAGE
2657
2695
  docs: >-
2658
- Type of Chat Event. There are six possible values:
2696
+ Type of Chat Event. There are eleven Chat Event types:
2659
2697
 
2698
+ - `SYSTEM_PROMPT`: The system prompt used to initialize the session.
2660
2699
 
2661
- - `SYSTEM_PROMPT`: Contains the system prompt for use in the session.
2700
+ - `CHAT_START_MESSAGE`: Marks the beginning of the chat session.
2662
2701
 
2702
+ - `USER_RECORDING_START_MESSAGE`: Marks when the client began streaming
2703
+ audio and the start of audio processing.
2663
2704
 
2664
- - `USER_MESSAGE`: Contains the message sent by the user.
2705
+ - `USER_MESSAGE`: A message sent by the user.
2665
2706
 
2707
+ - `USER_INTERRUPTION`: A user-initiated interruption while the assistant
2708
+ is speaking.
2666
2709
 
2667
- - `USER_INTERRUPTION`: Contains an interruption made by the user while the
2668
- agent is speaking.
2710
+ - `AGENT_MESSAGE`: A response generated by the assistant.
2669
2711
 
2712
+ - `FUNCTION_CALL`: A record of a tool invocation by the assistant.
2670
2713
 
2671
- - `AGENT_MESSAGE`: Contains the assistant’s message, generated by Hume’s
2672
- eLLM and supplemental LLM.
2714
+ - `FUNCTION_CALL_RESPONSE`: The result of a previously invoked function or
2715
+ tool.
2673
2716
 
2717
+ - `PAUSE_ONSET`: Marks when the client sent a `pause_assistant_message` to
2718
+ pause the assistant.
2674
2719
 
2675
- - `FUNCTION_CALL`: Contains the invocation of a tool.
2720
+ - `RESUME_ONSET`: Marks when the client sent a `resume_assistant_message`
2721
+ to resume the assistant.
2676
2722
 
2677
-
2678
- - `FUNCTION_CALL_RESPONSE`: Contains the tool response.
2723
+ - `CHAT_END_MESSAGE`: Indicates the end of the chat session.
2679
2724
  inline: true
2680
2725
  source:
2681
2726
  openapi: evi-openapi.json
@@ -2699,42 +2744,45 @@ types:
2699
2744
  The role of the entity which generated the Chat Event. There are four
2700
2745
  possible values:
2701
2746
 
2702
-
2703
2747
  - `USER`: The user, capable of sending user messages and
2704
2748
  interruptions.
2705
2749
 
2706
-
2707
2750
  - `AGENT`: The assistant, capable of sending agent messages.
2708
2751
 
2709
-
2710
2752
  - `SYSTEM`: The backend server, capable of transmitting errors.
2711
2753
 
2712
-
2713
2754
  - `TOOL`: The function calling mechanism.
2714
2755
  type:
2715
2756
  type: ReturnChatEventType
2716
2757
  docs: >-
2717
- Type of Chat Event. There are six possible values:
2718
-
2758
+ Type of Chat Event. There are eleven Chat Event types:
2719
2759
 
2720
- - `SYSTEM_PROMPT`: Contains the system prompt for use in the session.
2760
+ - `SYSTEM_PROMPT`: The system prompt used to initialize the session.
2721
2761
 
2762
+ - `CHAT_START_MESSAGE`: Marks the beginning of the chat session.
2722
2763
 
2723
- - `USER_MESSAGE`: Contains the message sent by the user.
2764
+ - `USER_RECORDING_START_MESSAGE`: Marks when the client began
2765
+ streaming audio and the start of audio processing.
2724
2766
 
2767
+ - `USER_MESSAGE`: A message sent by the user.
2725
2768
 
2726
- - `USER_INTERRUPTION`: Contains an interruption made by the user while
2727
- the agent is speaking.
2769
+ - `USER_INTERRUPTION`: A user-initiated interruption while the
2770
+ assistant is speaking.
2728
2771
 
2772
+ - `AGENT_MESSAGE`: A response generated by the assistant.
2729
2773
 
2730
- - `AGENT_MESSAGE`: Contains the assistant’s message, generated by
2731
- Hume’s eLLM and supplemental LLM.
2774
+ - `FUNCTION_CALL`: A record of a tool invocation by the assistant.
2732
2775
 
2776
+ - `FUNCTION_CALL_RESPONSE`: The result of a previously invoked
2777
+ function or tool.
2733
2778
 
2734
- - `FUNCTION_CALL`: Contains the invocation of a tool.
2779
+ - `PAUSE_ONSET`: Marks when the client sent a
2780
+ `pause_assistant_message` to pause the assistant.
2735
2781
 
2782
+ - `RESUME_ONSET`: Marks when the client sent a
2783
+ `resume_assistant_message` to resume the assistant.
2736
2784
 
2737
- - `FUNCTION_CALL_RESPONSE`: Contains the tool response.
2785
+ - `CHAT_END_MESSAGE`: Indicates the end of the chat session.
2738
2786
  message_text:
2739
2787
  type: optional<string>
2740
2788
  docs: >-
@@ -2886,6 +2934,7 @@ types:
2886
2934
  enum:
2887
2935
  - HUME_AI
2888
2936
  - CUSTOM_VOICE
2937
+ - OCTAVE_COMBINED
2889
2938
  source:
2890
2939
  openapi: evi-openapi.json
2891
2940
  VoiceId: