Skip to content

feat(regen): flux stt numerals and aura-2 multilingual tts voices#746

Open
GregHolmes wants to merge 7 commits into
mainfrom
gh/sdk-gen-2026-07-20
Open

feat(regen): flux stt numerals and aura-2 multilingual tts voices#746
GregHolmes wants to merge 7 commits into
mainfrom
gh/sdk-gen-2026-07-20

Conversation

@GregHolmes

@GregHolmes GregHolmes commented Jul 20, 2026

Copy link
Copy Markdown
Contributor

SDK regeneration for 2026-07-20 (Fern generator 5.14.18, spec ff8fd2b7), bumping the package to 7.5.1.

Features

  • Flux STT numerals — new types/listen_v2numerals.py (Literal["true","false"]) plus a numerals query param on listen.v2.connect() and the raw client. Connection-time only. (docs deepgram-docs#1020)
    • Wire coverage added: tests/custom/test_listen_v2_connect_wire.py (sync + async) pins that /v2/listen serializes numerals=true and omits it when absent — the /v2/listen handshake previously had no wire test.
    • Demonstrated in examples/14-transcription-live-websocket-v2.py.
  • New Aura-2 multilingual TTS voices — ~40 voices added to SpeakV1Model and AudioGenerateRequestModel (es/de/nl/fr/it/ja). Additive only; not breaking.

Breaking change from spec — mitigated (stays non-breaking)

  • The spec removed stt_latency from AgentV1LatencyReport (deepgram-docs#1006). AgentV1LatencyReport is a server-emitted (read-only) message, so stt_latency is re-added by hand as a read-side backward-compat field (frozen in .fernignore): report.stt_latency keeps resolving to None instead of raising AttributeError, with no request/wire impact. Covered by tests/custom/test_latency_report_stt_compat.py. Keeps this a minor/patch release, not a major.

Manual patches reconciled

All 27 previously-frozen patches were still needed and re-applied (byte-identical to main; the behavioral files don't appear in the net diff). Intended changes to frozen files: client_wrapper.py version annotations -> 7.5.1, and root + types __init__.py gained the new ListenV2Numerals export alongside the legacy aliases. New frozen entries: the two agent_v1latency_report.py shims and the two new custom tests. All .bak files deleted.

Verification

  • mypy src/deepgram — clean (857 files)
  • pytest — 331 passed, 1 skipped
  • ruff check src/deepgram — clean (pre-existing warnings in hand-maintained tests/manual + tests/custom only, identical on main)

fern-api Bot and others added 4 commits July 20, 2026 09:26
The 2026-07-20 regen dropped stt_latency from AgentV1LatencyReport after the
API spec removed it (deepgram-docs #1006). LatencyReport is a server-emitted
(read-only) message, so re-add the optional field by hand and freeze it: this
keeps report.stt_latency resolving to None instead of raising AttributeError,
avoiding a major-version break. No request/wire impact — the client never sends
a LatencyReport.
Adds tests/custom/test_listen_v2_connect_wire.py (sync + async), pinning that
listen.v2.connect() targets /v2/listen and serializes the new numerals param to
the wire as numerals=true (and omits it when absent) — the /v2/listen handshake
previously had no wire coverage. Demonstrates numerals in the v2 live example
and fixes a pre-existing f-string-without-placeholder lint there.
@GregHolmes GregHolmes changed the title chore: SDK regeneration 2026-07-20 feat(regen): Flux STT numerals and Aura-2 multilingual TTS voices Jul 20, 2026
@GregHolmes GregHolmes changed the title feat(regen): Flux STT numerals and Aura-2 multilingual TTS voices feat(regen): flux stt numerals and aura-2 multilingual tts voices Jul 20, 2026
@GregHolmes GregHolmes self-assigned this Jul 20, 2026
…2.connect

reference.md is frozen in .fernignore, so the Fern regen never adds these to
the listen.v2.connect() section — it jumped straight from keyterm to mip_opt_out.
Add the three missing params (in connect() signature order) by hand, incl. the
new numerals param from this regen (docs deepgram-docs#1020).
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants