Repository navigation
feat(regen): add usage redaction, Flux volume controls, and voices - #828
Merged
Merged
Conversation
GregHolmes
requested review from
deepgram-kiley and
dg-coreylweathers
as code owners
November 12, 2026 12:58
Contributor
|
GregHolmes
changed the title
GregHolmes
changed the title
dg-coreylweathers
approved these changes
Nov 12, 2026
dg-coreylweathers
left a comment
dg-coreylweathers
left a comment
Contributor
There was a problem hiding this comment.
What this PR does: This adds three new opt-in features to the Python SDK: a setting called redact_usage that keeps certain transcript keywords out of Deepgram's internal usage records, a setting called normalize_volume that controls loudness on text-to-speech audio, and a large batch of new voices across 20+ languages for the Flux text-to-speech engine.
What I checked:
- Is anything here a breaking change for existing users? No — everything added is optional and defaults to off, so code that worked before this PR keeps working exactly the same way.
- Did the SDK's existing hand-patched compatibility fixes survive this regeneration? Yes, including one that had accidentally been dropped once before in an earlier regeneration — I checked it specifically and it's still there.
- Do these new settings actually work against Deepgram's live servers, not just in theory? Yes. I connected directly to the production server and confirmed redact_usage behaves exactly as documented: it starts off, can be turned on mid-conversation, and once on it can't be turned back off for that session. I also confirmed one of the new voices (flux-ambreen-ur) genuinely exists and produces real audio on production today.
- Does it pass our tests? Yes — 1175 tests passed with nothing new broken. The PR's own notes mentioned some test failures caused by a local networking quirk; those didn't show up at all once I ran the tests with the correct setup, so that's not a real gap.
What to fix / what still needs you: Nothing blocking. One small cleanup: the PR's description shows an example using the text "true" in quotes for the redact_usage setting, but it's actually a true/false value, not text — it happens to work by accident today, but writing it as a plain true (no quotes) would be clearer for anyone copying the example.
GregHolmes
added a commit
that referenced
this pull request
Nov 13, 2026
🤖 I have created a release *beep* *boop* --- ## [7.19.0](v7.18.0...v7.19.0) (2026-11-12) ### Features * **Listen:** Add the optional `redact_usage` control to Listen V1 and V2 clients. Active Flux sessions can update it through `ListenV2Configure`, and `ListenV2ConfigureSuccess` reports the active setting. ([#828](#828)) ([0487431](0487431)) * **Flux TTS:** Add the `normalize_volume` control for batch and WebSocket synthesis. Expand the typed Flux voice catalog with new multilingual voices, including support for Deepgram Agent TTS providers. ([#828](#828)) ([0487431](0487431)) ## Highlights - Control usage redaction for Listen V1 and V2 connections; Flux users can update the setting during an active session and inspect the server-confirmed value. - Control Flux speech-volume normalization for batch and streaming synthesis. - Select from the expanded typed catalog of multilingual Flux voices, including in Deepgram Agent TTS configuration. --- This PR was prepared for the Deepgram Python SDK release workflow.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
fernapi/fern-python-sdk5.39.3.redact_usagecontrols, Speak V2normalize_volume, and an expanded typed Flux voice catalog.redact_usagewhile restoring dict-style Listen V2 response access.7.18.0; Release Please will create the next feature-release version.What This Enables
redact_usagewhen opening Listen V1 or V2 connections to control usage redaction, then update an active Flux session withconnection.send_configure(ListenV2Configure(redact_usage=True)). The server reports the active setting inListenV2ConfigureSuccess.redact_usage.normalize_volume=TrueorFalseto Flux batch synthesis withclient.speak.v2.audio.generate(...), or use the string form ("true"/"false") onclient.speak.v2.connect(...). Volume normalization is enabled by default; opting out still applies peak limiting.DeepgramModelcatalog for multilingual Flux voices such asflux-ambreen-ur, including when configuring the Deepgram Agent TTS provider.with_raw_response()clients.Validation
redact_usage, Listen V2 Configure acknowledgements, and Speak V2normalize_volumewithflux-ambreen-ur.poetry run mypy src/ tests/typecheckpoetry run ruff check .retains 57 existing violations in examples, manual tests, andtests/custom/test_transport.py.poetry run pytest -rP .: 1,102 passed, 4 skipped, 1 expected failure. The remaining 53 WireMock failures are allConnection refusedbecause Docker Compose reports the dynamic WireMock mapping aslocalhost:0.