Dhruv Kant Ladia 08b8dfb131 feat(sarvam): add glm5.2, gemma4 and sarvam-105b-conversations model support (#6841)

2026-08-24 13:33:54 -04:00

..

test_interruption

fix: preserve adaptive interruption across tool calls (#6564)

2026-08-14 13:48:24 +01:00

test_realtime

fix(openai realtime): recover when the server anchors to a deleted item (#6964)

2026-08-24 16:03:21 +08:00

test_utils

fix(audio): check AudioArrayBuffer frame size after resampling (#6879)

2026-08-17 15:00:13 +08:00

utils

…

__init__.py

…

.gitattributes

…

.gitignore

…

change-sophie.opus

…

change-sophie.wav

…

concurrency.py

…

conftest.py

…

docker-compose.yml

…

Dockerfile.tests

…

Dockerfile.toxiproxy

…

fake_io.py

…

fake_llm.py

…

fake_realtime.py

fix(voice): agent and user state while a tool runs (#6937)

2026-08-21 15:37:49 +08:00

fake_session.py

…

fake_stt.py

…

fake_tts.py

…

fake_turn_detector_ws.py

…

fake_turn_stt.py

…

fake_vad.py

…

lk_server.py

…

long_synthesize_multi_tokens.txt

…

long_synthesize.txt

…

long_transcript.txt

…

long.mp3

…

Makefile

…

test_agent_session.py

fix(voice): agent and user state while a tool runs (#6937)

2026-08-21 15:37:49 +08:00

test_agent_task_close_race.py

…

test_agent_update_options.py

…

test_aio_itertools.py

…

test_aio.py

…

test_aligned_transcript_fallback.py

fix(voice): keep the turn transcript when the tts sends no word timings (#6704)

2026-08-17 11:34:19 +08:00

test_amd_classifier.py

…

test_async_mcp.py

…

test_audio_decoder.py

…

test_audio_emitter.py

…

test_audio_recognition_aclose.py

fix: wait for final user turn during close (#6897)

2026-08-19 10:35:10 +01:00

test_audio_recognition_handoff.py

…

test_audio_recognition_push_audio.py

…

test_audio_recognition_turn_detection.py

fix(voice): export uncommitted user_turn spans on close (#6592)

2026-08-12 10:16:51 +01:00

test_audio_sink_proxy.py

fix(voice): place recorded audio where it actually played (#6921)

2026-08-24 16:04:49 +08:00

test_audio_turn_detector_fallback.py

feat: correlate inference requests with session ids (#6863)

2026-08-18 11:01:42 +01:00

test_chat_ctx.py

…

test_chat_item_proto_coverage.py

fix(proto): one ChatItem serializer that carries every declared field (#6756)

2026-08-20 13:49:54 -04:00

test_cli_log_level.py

…

test_config.py

…

test_connection_pool.py

fix: respect non-retryable STT errors (#6759)

2026-08-09 17:54:32 +01:00

test_convert_html_docs.py

…

test_debounce.py

…

test_dob.py

…

test_drain_timeout.py

…

test_end_of_turn_metrics.py

…

test_endpointing.py

…

test_evals.py

…

test_exceptions.py

fix(agents): omit separator for empty exception details (#6894)

2026-08-18 15:04:30 -04:00

test_expr_markup.py

…

test_expressive_toggle.py

feat(expressive): allow dynamic updates via Agent and update_options (#6886)

2026-08-17 19:43:24 -04:00

test_false_interruption_resume.py

fix: avoid duplicate agent speech end (#6938)

2026-08-21 17:30:35 +01:00

test_filler.py

…

test_foreground_run_tracking.py

…

test_google_credentials.py

fix(google): allow explicit project for Speech-to-Text (#6810)

2026-08-12 10:59:14 +08:00

test_google_thought_signatures.py

…

test_http_context_helper.py

fix: raise the shared http session total timeout to 900s (#6941)

2026-08-24 10:54:40 +01:00

test_inference_avatar.py

…

test_inference_llm_retry.py

feat: correlate inference requests with session ids (#6863)

2026-08-18 11:01:42 +01:00

test_inference_llm_usage.py

fix(inference): don't crash the session on null usage token counts (#6763)

2026-08-10 16:48:02 +08:00

test_inference_stt_aligned_transcript_claim.py

…

test_inference_stt_context.py

…

test_inference_stt_fallback.py

…

test_inference_stt_start_of_speech.py

feat: correlate inference requests with session ids (#6863)

2026-08-18 11:01:42 +01:00

test_inference_tts_alignment.py

feat: correlate inference requests with session ids (#6863)

2026-08-18 11:01:42 +01:00

test_inference_tts_fallback.py

…

test_inference_utils.py

feat: correlate inference requests with session ids (#6863)

2026-08-18 11:01:42 +01:00

test_interrupt_protected_speech.py

fix(voice): don't raise when interrupting a cancelled or finished speech (#6767)

2026-08-10 20:01:13 +08:00

test_interruption_hold_window.py

…

test_ipc.py

…

test_ivr_activity.py

…

test_judge.py

…

test_krisp_frame_buffer.py

…

test_langgraph.py

…

test_language.py

…

test_llm_fallback.py

fix: fallback adapters report the active instance's model/provider (#6690)

2026-08-10 10:19:47 +08:00

test_llm_node_metrics.py

…

test_llm_telemetry.py

fix(telemetry): report cached input tokens on LLM spans (#6852)

2026-08-17 16:14:40 +08:00

test_llm_utils.py

…

test_llm.py

…

test_metrics_usage.py

feat: expose cache_creation_tokens in LLM metrics (#6655)

2026-08-11 10:48:18 +08:00

test_mood.py

…

test_multi_speaker_adapter.py

…

test_multi_speaker_primary_speaker.py

…

test_nested_agent_task.py

feat(agents): serialize the inline AgentTasks awaited from one turn (#6862)

2026-08-21 09:08:32 -04:00

test_openai_realtime_chat_ctx.py

fix(openai realtime): recover when the server anchors to a deleted item (#6964)

2026-08-24 16:03:21 +08:00

test_personaplex_realtime_model.py

…

test_playback_progress.py

fix(voice): place recorded audio where it actually played (#6921)

2026-08-24 16:04:49 +08:00

test_playout_launch_pause.py

fix: hold pending replies at playout launch (#6745)

2026-08-11 12:36:51 +01:00

test_plugin_anthropic.py

…

test_plugin_assemblyai_stt.py

…

test_plugin_aws_stt.py

…

test_plugin_azure_stt.py

…

test_plugin_azure_tts.py

feat(azure): replace unknown tts model string with host/region + voice (#6775)

2026-08-10 22:44:12 +01:00

test_plugin_bland_tts.py

feat(bland): BTTS ws integration (#6752)

2026-08-20 12:33:12 -04:00

test_plugin_cartesia_tts.py

fix: respect non-retryable STT errors (#6759)

2026-08-09 17:54:32 +01:00

test_plugin_cerebras_request.py

…

test_plugin_cerebras.py

…

test_plugin_deepgram_stt.py

…

test_plugin_deepgram_tts.py

…

test_plugin_elevenlabs_stt.py

fix(elevenlabs): send keyterms on the realtime STT WebSocket (#6746)

2026-08-13 16:03:38 +08:00

test_plugin_elevenlabs_tts.py

fix(elevenlabs): stop routing TTS audio to an interrupted stream (#6936)

2026-08-24 16:30:46 +08:00

test_plugin_fishaudio_tts.py

…

test_plugin_gnani_stt.py

…

test_plugin_gnani_tts.py

…

test_plugin_google_gemini_tts.py

…

test_plugin_google_llm.py

…

test_plugin_google_realtime.py

fix(google): keep unspoken model text out of the realtime transcript (#6955)

2026-08-23 14:20:15 -07:00

test_plugin_google_stt.py

…

test_plugin_gradium_stt.py

…

test_plugin_inworld_tts.py

…

test_plugin_openai_realtime_reasoning.py

…

test_plugin_openai_responses.py

fix(inference): keep a stream retryable until it emits generation (#6697)

2026-08-11 07:54:22 -04:00

test_plugin_openai_stt_context.py

…

test_plugin_openai_tts.py

fix(openai): read TTS response by content-type, not model name (#6930)

2026-08-21 21:26:56 +08:00

test_plugin_openai_websocket_urls.py

…

test_plugin_perplexity_responses.py

…

test_plugin_perplexity.py

…

test_plugin_rime_tts.py

fix(rime): remove Arcana model support (#6773)

2026-08-20 12:14:19 -04:00

test_plugin_sarvam_llm.py

feat(sarvam): add glm5.2, gemma4 and sarvam-105b-conversations model support (#6841)

2026-08-24 13:33:54 -04:00

test_plugin_sarvam_stt_streaming.py

feat(sarvam): add realtime streaming STT (#6562)

2026-08-24 13:12:51 -04:00

test_plugin_soniox_stt.py

…

test_plugin_soniox_tts.py

…

test_plugin_speechify.py

feat(speechify): stream with timestamps via /v1/audio/stream/with-timestamps (#6926)

2026-08-21 14:35:17 -04:00

test_plugin_xai_tts.py

…

test_realtime_adaptive_interruption.py

fix: preserve adaptive interruption across tool calls (#6564)

2026-08-14 13:48:24 +01:00

test_realtime_agent_state_during_tool.py

fix(voice): agent and user state while a tool runs (#6937)

2026-08-21 15:37:49 +08:00

test_realtime_auto_disable_turn_detection.py

…

test_realtime_cancel_tool_reply.py

fix(voice): answer every tool call (#6785)

2026-08-14 15:03:42 +08:00

test_realtime_fallback.py

fix(llm): preserve provider events on restart (#6568)

2026-08-18 22:23:12 +08:00

test_realtime_input_speech_interrupt.py

…

test_realtime_interrupt_cancels_generation.py

fix(realtime): cancel the generation when a speech is interrupted (#6962)

2026-08-24 14:34:53 +08:00

test_realtime_message_metrics.py

…

test_realtime_reply_chat_ctx.py

…

test_realtime_tool_call_created_at.py

…

test_recorder_io.py

fix(voice): place recorded audio where it actually played (#6921)

2026-08-24 16:04:49 +08:00

test_recording.py

fix(voice): place recorded audio where it actually played (#6921)

2026-08-24 16:04:49 +08:00

test_remote_session.py

fix: stop dropping session host responses (#6781)

2026-08-10 22:40:37 -07:00

test_response_format.py

…

test_room_io.py

fix: don't count discarded room audio as played (#6728)

2026-08-10 22:17:34 +01:00

test_room.py

…

test_run_input_errors.py

fix: treat a silent agent turn as silence, not a run_input error (#6779)

2026-08-10 20:40:38 -07:00

test_running_tool_placeholders.py

…

test_schema_gemini.py

…

test_session_close_user_turn.py

fix: wait for final user turn during close (#6897)

2026-08-19 10:35:10 +01:00

test_session_host.py

fix(proto): one ChatItem serializer that carries every declared field (#6756)

2026-08-20 13:49:54 -04:00

test_speaker_id_grouping.py

…

test_speech_handle_exception.py

…

test_speech_start_time_persistence.py

…

test_start_capture_run.py

fix(voice): capture the full on_enter output when starting with capture_run (#6872)

2026-08-17 14:57:38 +08:00

test_stt_base.py

fix: respect non-retryable STT errors (#6759)

2026-08-09 17:54:32 +01:00

test_stt_context.py

…

test_stt_fallback.py

fix: fallback adapters report the active instance's model/provider (#6690)

2026-08-10 10:19:47 +08:00

test_stt_prewarm.py

…

test_stt.py

…

test_supervised_proc_memory.py

…

test_telemetry_exception_redaction.py

feat(telemetry): tag content-bearing trace and log keys with lk.pii. for redaction AGT-3074 (#6356)

2026-08-20 17:44:07 +01:00

test_telemetry_recording_disabled.py

…

test_tokenize_sentence_spans.py

fix(tokenize): include the last character in the final sentence span (#6866)

2026-08-17 15:54:36 +08:00

test_tokenizer_xml_markup.py

…

test_tokenizer.py

…

test_tool_output_per_call.py

fix(voice): answer every tool call (#6785)

2026-08-14 15:03:42 +08:00

test_tool_proxy.py

…

test_tool_results_preserved_on_interruption.py

fix(voice): answer every tool call (#6785)

2026-08-14 15:03:42 +08:00

test_tool_search.py

…

test_tools.py

fix(voice): agent and user state while a tool runs (#6937)

2026-08-21 15:37:49 +08:00

test_trace_types_pii.py

feat(telemetry): tag content-bearing trace and log keys with lk.pii. for redaction AGT-3074 (#6356)

2026-08-20 17:44:07 +01:00

test_transcript_sync_markup.py

fix(voice): keep transcript pacing speed nonzero for syllable-free segments (#6751)

2026-08-17 12:05:30 +08:00

test_transcription_filter.py

…

test_transcription_timeout.py

…

test_tts_fallback.py

fix: fallback adapters report the active instance's model/provider (#6690)

2026-08-10 10:19:47 +08:00

test_tts.py

fix(rime): remove Arcana model support (#6773)

2026-08-20 12:14:19 -04:00

test_turn_detection_cloud_stream.py

…

test_turn_detection_fsm.py

feat: correlate inference requests with session ids (#6863)

2026-08-18 11:01:42 +01:00

test_update_agent_long_on_enter.py

…

test_user_turn_exceeded.py

…

test_utils_env.py

…

test_vad.py

fix(silero): update bundled VAD checkpoint to v6.2 (#6816)

2026-08-12 17:07:54 +08:00

test_wait_for_idle_cancellation.py

fix: prevent wait_for_idle cancellation from poisoning user turns (#6913)

2026-08-20 16:21:59 +08:00

test_worker_config.py

…

test_worker_connection.py

…

test_workflows.py

test: stabilize text email collection eval (#6791)

2026-08-11 15:28:44 +01:00

toxic_proxy.py

…

virtual_time.py

…