Releases: HumeAI/hume-python-sdk
v0.7.8
Introducing Octave TTS
This release includes support for Hume AI's Octave TTS API. Learn more on our docs page, or try it out on the web at hume.ai.
Unlike conventional TTS that merely "reads" words, Octave is a speech-language model that understands what words mean in context, unlocking a new level of expressiveness. It acts out characters, generates voices from prompts, and takes instructions to modify the emotion and style of a given utterance.
What's Changed
- 🌿 Fern Regeneration -- February 26, 2025 by @fern-api in #236
- Introduce
AsyncHumeClient.tts
client for Octave. - (fix, breaking) Renamed method
create_prompt_verison
tocreate_prompt_version
.
- Introduce
Full Changelog: v0.7.7...v0.7.8
v0.7.7
What's Changed
- 🌿 Fern Regeneration -- January 15, 2025 by @fern-api in #222
- Add
WebhookEvent
payload types.
- Add
- 🌿 Fern Regeneration -- January 15, 2025 by @fern-api in #224
- Updates to latest fern generator.
HttpClient
now takes callables forbase_headers
,base_timeout
, andbase_url
instead of hardcoded values
Full Changelog: v0.7.6...v0.7.7
v0.7.6
What's Changed
- 🌿 Fern Regeneration -- January 13, 2025 by @fern-api in #220
- Add support for
.webhooks
to theConfig
resource - Add support for
.config_id
filter to the.empathic_voice.chats.listChats(...)
method - Add support for
.most_recent_config
to theChatGroup
resource.
- Add support for
Full Changelog: v0.7.5...v0.7.6
v0.7.5
What's Changed
- 🌿 Fern Regeneration -- November 6, 2024 by @fern-api in #211
- Adds
interim
field toUserMessage
events in EVI chats, to indicate whether a user message is "interim". - Adds enum members for Claude Haiku 3.5.
- Adds
- 🌿 Fern Regeneration -- December 4, 2024 by @fern-api in #217
⚠️ (break) Updated voice parameters: removedarticulation
- Improved type of
SessionSettings.variables
Full Changelog: v0.7.4...v0.7.5
v0.7.4
What's Changed
- 🌿 Fern Regeneration -- October 31, 2024 by @fern-api in #210
- (fix) added
CUSTOM_VOICE
as a member to theReturnVoice.voice_provider
enum - (fix) added correct return types for the audio reconstruction endpoints
- (feat) Typed
.language_model.model_resource
onPOST /v0/evi/configs
as an enum instead of a string.
- (fix) added
Full Changelog: v0.7.3...v0.7.4
v0.7.3
What's Changed
What's Changed
- Add support for
hume.empathic_voice.chats.get_audio(...)
for getting a URL that you can use to download the reconstructed audio for a chat- Note: this method will not return a ready-to-use URL right away, it will kick off a job to ready the URL. You can use this like
result = await hume.empathic_voice.chats.get_audio(...) await asyncio.sleep(10) result = await hume.empathic_voice.chats.get_audio(...) if result.status != "COMPLETE": raise TimeoutError('Job did not complete in 10 seconds') url_containing_download = json.signed_audio_url
- Note: this method will not return a ready-to-use URL right away, it will kick off a job to ready the URL. You can use this like
- Add support for
hume.empathic_voice.chat_groups.get_audio(...)
to list urls containing reconstructed audio for many chats in a chat group - Custom Voices
- Add support for new base voice
SUNNY
- Updates custom voice parameters to reflect the latest parameters.
- Add support for new base voice
Full Changelog: v0.7.2...v0.7.3
hume 0.7.2
This release fixes a validation error in ReturnConfig
when creating a config with a null custom_voice
. Previously, this would raise a pydantic ValidationError
. Now, null values for custom_voice
are properly handled.
Example of fixed usage:
client = AsyncHumeClient(api_key=<HUME_API_KEY>) # replace <HUME_API_KEY> with your API key
resp = await client.empathic_voice.configs.create_config(name="test", evi_version=1)
# This now works without raising a ValidationError, even if custom_voice is implicitly None
hume 0.7.1
This major release introduces significant architectural changes to improve SDK functionality and developer experience.
Highlights:
- New base clients:
AsyncHumeClient
andHumeClient
- Refactored API structure with distinct modules for Expression Measurement and Empathic Voice Interface
- Enhanced type safety and asynchronous support
- More granular client configuration options
- Legacy support maintained for backward compatibility
Please refer to the Migration Guide for detailed information on breaking changes and how to update your code.
We recommend all users upgrade to this version to leverage the latest improvements and prepare for future updates.
hume 0.7.0
This release was buggy and therefore previously yanked; see v0.7.1
.
0.7.0-rc2
Publish pre-release version of auto-generated SDK