c261d08036
feat: require complete KB-grounded answers (no dropped facts) + ask did-that-answer-your-question after each answer, per ai-operator spec
main
a.arystanbek2026-08-30 21:22:54 +00:00
78dfe6a9e1
fix: do not fire no-speech reprompt while caller is still mid-utterance with live partial ASR
a.arystanbek2026-08-30 21:19:02 +00:00
f978702a6c
fix: retry voice turn once on Postgres deadlock instead of failing the call with a technical-error handoff
a.arystanbek2026-08-30 21:14:25 +00:00
f82f27c035
fix: gpt-5-nano too unreliable for aimaq voice decisions, switch to gpt-5-mini
a.arystanbek2026-08-30 21:04:55 +00:00
75d105f636
fix: support GPT-5 request contract (max_completion_tokens, reasoning_effort) and switch aimaq voice decisions to gpt-5-nano
a.arystanbek2026-08-30 20:37:35 +00:00
9bf367abf4
fix: filter RU/KZ stopwords from KB search so filler words can't false-match
didar2026-08-31 01:08:11 +05:00
3c5c233071
Merge pull request 'fix: stop no-answer retry racing dialplan hangup + fix event catalog validation' (#11) from fix/retry-channel-resolution-and-event-catalog into main
arystanbek2026-08-30 20:04:04 +00:00
9fdaa9472f
Revert "fix: let voice early-plan turns answer from the FAQ knowledge base"
didar2026-08-31 00:54:38 +05:00
1dc37d2764
Merge pull request 'fix: treat AST_CAUSE no-route/unallocated as a no-answer retry outcome' (#10) from fix/no-route-retry-cause into main
arystanbek2026-08-30 19:43:43 +00:00
8ced7a59e3
fix: let voice early-plan turns answer from the FAQ knowledge base
didar2026-08-31 00:38:06 +05:00
d2438b6954
feat: canonical intent taxonomy for AI operator (kb_answer -> intent_code)
didar2026-08-31 00:17:51 +05:00
3826f5704a
Merge pull request 'fix: track escalation for legacy AI voice handoff path' (#9) from fix/handoff-escalation-tracking into main
arystanbek2026-08-30 19:08:20 +00:00
ff85fa27fb
Merge pull request 'feat: aimaq voice AI persona Zhanna + language-choice greeting' (#7) from feature/aimaq-zhanna-persona-language-choice into main
arystanbek2026-08-30 07:22:16 +00:00
9e4c47eddd
feat: enhance AudioSocketMediaRuntime to skip filler acks for closing intents and throttle repeated filler acks
didar2026-08-30 11:30:55 +05:00
a013ac95e8
Merge pull request 'fix: voice AI spells out numbers/dates for TTS' (#5) from fix/voice-number-pronunciation into main
arystanbek2026-08-30 06:22:05 +00:00
92095ff5d6
fix: release reserved L2 agent when AMI redirect fails during escalation
Hermes Agent2026-08-29 14:01:40 +05:00
8c131440fc
config: route voice_lab_ai / ivr_aimaq_ai_ru queues through L2 agent pool
Hermes Agent2026-08-29 13:37:30 +05:00
6ef2e6b81f
fix: update ElevenLabs voice IDs for Russian and Kazakh languages
didar2026-08-29 13:36:27 +05:00
16ce9659db
Merge pull request 'feat: L1->L2 agent pool and routing engine for voice escalation' (#4) from feature/l1-l2-routing-engine into main
arystanbek2026-08-29 08:31:40 +00:00
1ad4e1ec6e
feat: surface L2 agent pool and escalations on supervisor screen
Hermes Agent2026-08-29 13:28:14 +05:00
a651b9c086
fix: add name correction action in _voice_downstream_name_update function
didar2026-08-29 01:53:58 +05:00
78007dae38
fix: improve name matching logic in _voice_reply_with_name function
didar2026-08-29 01:32:08 +05:00
d2a9df36d1
feat: add cache_fingerprint method to TTSProvider and its subclasses for voice/model configuration
didar2026-08-29 01:23:23 +05:00
c590528694
feat: add debug audio dump functionality for ElevenLabs ASR provider
didar2026-08-23 13:38:56 +05:00
9a09a39849
feat: enhance ElevenLabs ASR provider to accumulate autonomous VAD commits
didar2026-08-23 13:20:31 +05:00
206b0bafe0
feat: update real-time commit strategy for ElevenLabs ASR to VAD
didar2026-08-23 13:05:58 +05:00
1a91dc4489
feat: add tenant ID handling for sales voice and telegram sync requests
didar2026-08-23 12:55:25 +05:00
8dd3e238f2
feat: enhance low signal transcript handling with finalization checks and logging
didar2026-08-23 12:40:59 +05:00
642104ec0d
Merge pull request 'feat: make ai_operator_settings config code the source of truth' (#3) from feat/ai-operator-config-source-of-truth into main
didar2026-08-21 16:17:12 +00:00
c5dd87ee32
feat: make ai_operator_settings config code the source of truth
didarandClaude Sonnet 52026-08-21 21:00:34 +05:00
6d9b0ce98a
feat: add aimaq production env file (TTS/ASR/LLM config in git)
Codex2026-08-21 13:37:25 +00:00
fae68aaaaf
ci: FINAL test ? Gitea 1.27.2 + runner v3.2.0 + web UI fixed
Codex2026-08-21 07:11:50 +00:00
ca31a867a9
ci: FINAL test ? Gitea 1.27.2 + runner v3.2.0 + web UI fixed
Codex2026-08-21 07:10:54 +00:00
9c3cee24e9
ci: test pipeline end-to-end after runner fix
Codex2026-08-21 06:45:25 +00:00
ffa779ffa0
Move ai-voice-runtime-service host port off 9019 to avoid cross-tenant conflict
didar2026-08-20 23:22:16 +05:00
13ef87eb2c
Fix voice greeting test marker to hit the actually-served default
didar2026-08-20 22:53:27 +05:00
c5e11fde3d
Add test marker to voice greeting to verify Gitea CI/CD deploy
didar2026-08-20 22:47:34 +05:00
4e1039ab58
Merge pull request 'Play a filler ack during slow voice decisions and cache KB search rows' (#2) from voice-latency-llm-tts-improvements into main
didar2026-08-20 09:47:45 +00:00
9cc5d802d6
Merge pull request 'Add architecture longread and remove dead code found during review' (#1) from worktree-call-center-review into main
didar2026-08-20 09:42:15 +00:00
a.arystanbek
didar
arystanbek