Skip to content

Agents-assignment_MitaliMehta #476

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Mitalimehta02 wants to merge 5 commits into
base: main
Choose a base branch
from
Open

Agents-assignment_MitaliMehta #476

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
4b3decb
Create interrupt_handle.py
Mitalimehta02 Feb 2, 2026
152f17a
Create interrupt_config.py
Mitalimehta02 Feb 2, 2026
28dcf58
Create Proof-of-working
Mitalimehta02 Feb 2, 2026
029ffbb
Update Proof-of-working
Mitalimehta02 Feb 2, 2026
70e06cb
Add files via upload
Mitalimehta02 Feb 2, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
136 changes: 136 additions & 0 deletions Proof-of-working
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,136 @@
2025-12-04 16:43:48,850 INFO livekit livekit_ffi::server:139:livekit_ffi::server - initializing ffi server v0.12.40
2025-12-04 16:43:49,249 INFO livekit.agents starting worker
2025-12-04 16:43:49,249 INFO livekit.agents starting inference executor
2025-12-04 16:43:49,273 INFO livekit.agents initializing process
2025-12-04 16:43:53,941 INFO livekit.agents process initialized
2025-12-04 16:43:53,948 INFO livekit.agents HTTP server listening on :63063
2025-12-04 16:43:53,966 INFO livekit.agents initializing job runner
2025-12-04 16:43:53,969 INFO basic-agent [PREWARM] loading VAD model
2025-12-04 16:43:54,138 INFO basic-agent [PREWARM] VAD model loaded
2025-12-04 16:43:54,140 INFO livekit.agents job runner initialized
2025-12-04 16:43:54,185 INFO basic-agent [ENTRYPOINT] starting AgentSession for room=mock_room
2025-12-04 16:43:54,669 INFO basic-agent [AGENT] on_enter -> generating initial reply
2025-12-04 16:43:55,159 INFO basic-agent [STATE] agent_state changed: initializing -> listening | user_state=listening
2025-12-04 16:43:55,166 WARNING livekit.agents resume_false_interruption is enabled but audio output does not support pause, it will be ignored
2025-12-04 16:43:55,175 INFO basic-agent [STATE] agent_state changed: listening -> thinking | user_state=listening
2025-12-04 16:43:56,488 INFO basic-agent [TOOL] Looking up weather for New York
2025-12-04 16:43:56,497 INFO livekit.agents LLM metrics
2025-12-04 16:43:57,870 INFO livekit.agents LLM metrics
2025-12-04 16:43:59,578 INFO basic-agent [STATE] agent_state changed: thinking -> speaking | user_state=listening
2025-12-04 16:44:00,676 INFO livekit.agents STT metrics
2025-12-04 16:44:00,846 INFO livekit.agents TTS metrics
2025-12-04 16:44:03,698 INFO livekit.agents STT metrics
2025-12-04 16:44:04,538 INFO basic-agent [STATE] agent_state changed: speaking -> listening | user_state=listening
2025-12-04 16:44:06,648 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=listening
2025-12-04 16:44:07,104 INFO livekit.agents STT metrics
2025-12-04 16:44:07,346 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=listening
2025-12-04 16:44:07,386 INFO livekit.agents STT metrics
2025-12-04 16:44:07,386 INFO basic-agent [STT] FINAL text='Yeah.' | agent_state=listening | user_state=listening
2025-12-04 16:44:07,395 INFO basic-agent [LOGIC] words=['yeah'] | is_soft=True | has_hard=False | agent_state=listening
2025-12-04 16:44:07,396 INFO basic-agent [LOGIC] SOFT input while agent NOT speaking -> normal user turn (agent will respond)
2025-12-04 16:44:08,888 INFO livekit.agents LLM metrics
2025-12-04 16:44:09,836 INFO livekit.agents EOU metrics
2025-12-04 16:44:09,849 INFO basic-agent [STATE] agent_state changed: listening -> thinking | user_state=listening
2025-12-04 16:44:09,849 INFO basic-agent [STATE] agent_state changed: thinking -> speaking | user_state=listening
2025-12-04 16:44:11,146 INFO livekit.agents TTS metrics
2025-12-04 16:44:12,376 INFO livekit.agents STT metrics
2025-12-04 16:44:15,043 INFO basic-agent [STATE] agent_state changed: speaking -> listening | user_state=listening
2025-12-04 16:44:15,418 INFO livekit.agents STT metrics
2025-12-04 16:44:16,336 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=listening
2025-12-04 16:44:16,988 INFO basic-agent [STT] PARTIAL text='There are no' | agent_state=listening | user_state=speaking
2025-12-04 16:44:17,741 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=listening
2025-12-04 16:44:17,990 INFO basic-agent [STT] PARTIAL text='Tell me about volcanoes.' | agent_state=listening | user_state=listening
2025-12-04 16:44:17,998 INFO livekit.agents STT metrics
2025-12-04 16:44:18,006 INFO basic-agent [STT] FINAL text='Tell me about volcanoes.' | agent_state=listening | user_state=listening
2025-12-04 16:44:18,008 INFO basic-agent [LOGIC] words=['tell', 'me', 'about', 'volcanoes'] | is_soft=False | has_hard=False | agent_state=listening
2025-12-04 16:44:18,010 INFO basic-agent [LOGIC] NORMAL input while agent NOT speaking -> normal user turn
2025-12-04 16:44:18,098 INFO livekit.agents EOU metrics
2025-12-04 16:44:18,098 INFO basic-agent [STATE] agent_state changed: listening -> thinking | user_state=listening
2025-12-04 16:44:19,746 INFO livekit.agents LLM metrics
2025-12-04 16:44:20,470 INFO basic-agent [STATE] agent_state changed: thinking -> speaking | user_state=listening
2025-12-04 16:44:23,006 INFO livekit.agents STT metrics
2025-12-04 16:44:23,935 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=speaking
2025-12-04 16:44:24,586 INFO livekit.agents STT metrics
2025-12-04 16:44:24,590 INFO basic-agent [STT] FINAL text='Okay.' | agent_state=speaking | user_state=speaking
2025-12-04 16:44:24,595 INFO basic-agent [LOGIC] words=['okay'] | is_soft=True | has_hard=False | agent_state=speaking
2025-12-04 16:44:24,598 INFO basic-agent [LOGIC] SOFT backchannel while speaking -> IGNORE COMPLETELY (no interrupt, no new turn)
2025-12-04 16:44:24,645 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=speaking
2025-12-04 16:44:24,726 WARNING livekit.agents skipping reply to user input, current speech generation cannot be interrupted
2025-12-04 16:44:25,338 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=speaking
2025-12-04 16:44:26,138 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=speaking
2025-12-04 16:44:26,188 INFO livekit.agents STT metrics
2025-12-04 16:44:26,188 INFO basic-agent [STT] FINAL text='Yeah.' | agent_state=speaking | user_state=listening
2025-12-04 16:44:26,196 INFO basic-agent [LOGIC] words=['yeah'] | is_soft=True | has_hard=False | agent_state=speaking
2025-12-04 16:44:26,197 INFO basic-agent [LOGIC] SOFT backchannel while speaking -> IGNORE COMPLETELY (no interrupt, no new turn)
2025-12-04 16:44:26,268 WARNING livekit.agents skipping reply to user input, current speech generation cannot be interrupted
2025-12-04 16:44:26,636 INFO livekit.agents TTS metrics
2025-12-04 16:44:27,246 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=speaking
2025-12-04 16:44:27,926 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=speaking
2025-12-04 16:44:28,097 INFO livekit.agents STT metrics
2025-12-04 16:44:28,097 INFO basic-agent [STT] FINAL text='Right.' | agent_state=speaking | user_state=listening
2025-12-04 16:44:28,107 INFO basic-agent [LOGIC] words=['right'] | is_soft=True | has_hard=False | agent_state=speaking
2025-12-04 16:44:28,107 INFO basic-agent [LOGIC] SOFT backchannel while speaking -> IGNORE COMPLETELY (no interrupt, no new turn)
2025-12-04 16:44:28,186 WARNING livekit.agents skipping reply to user input, current speech generation cannot be interrupted
2025-12-04 16:44:29,848 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=speaking
2025-12-04 16:44:30,496 INFO livekit.agents STT metrics
2025-12-04 16:44:30,501 INFO basic-agent [STT] FINAL text='Stop.' | agent_state=speaking | user_state=speaking
2025-12-04 16:44:30,503 INFO basic-agent [LOGIC] words=['stop'] | is_soft=False | has_hard=True | agent_state=speaking
2025-12-04 16:44:30,506 INFO basic-agent [LOGIC] HARD interrupt while speaking -> calling session.interrupt(force=True)
2025-12-04 16:44:30,515 INFO basic-agent [STATE] agent_state changed: speaking -> listening | user_state=speaking
2025-12-04 16:44:30,539 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=listening
2025-12-04 16:44:30,686 WARNING livekit.agents preemptive generation enabled but chat context or tools have changed after `on_user_turn_completed`
2025-12-04 16:44:30,690 INFO livekit.agents EOU metrics
2025-12-04 16:44:30,701 INFO basic-agent [STATE] agent_state changed: listening -> thinking | user_state=listening
2025-12-04 16:44:31,716 INFO livekit.agents LLM metrics
2025-12-04 16:44:33,675 INFO basic-agent [STATE] agent_state changed: thinking -> speaking | user_state=listening
2025-12-04 16:44:33,679 INFO livekit.agents TTS metrics
2025-12-04 16:44:35,478 INFO livekit.agents STT metrics
2025-12-04 16:44:36,346 INFO basic-agent [STATE] agent_state changed: speaking -> listening | user_state=listening
2025-12-04 16:44:38,046 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=listening
2025-12-04 16:44:38,676 INFO basic-agent [STT] PARTIAL text='Daniel' | agent_state=listening | user_state=speaking
2025-12-04 16:44:39,236 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=listening
2025-12-04 16:44:39,476 INFO livekit.agents STT metrics
2025-12-04 16:44:39,492 INFO basic-agent [STT] FINAL text='Ten more trucks.' | agent_state=listening | user_state=listening
2025-12-04 16:44:39,496 INFO basic-agent [LOGIC] words=['ten', 'more', 'trucks'] | is_soft=False | has_hard=False | agent_state=listening
2025-12-04 16:44:39,496 INFO basic-agent [LOGIC] NORMAL input while agent NOT speaking -> normal user turn
2025-12-04 16:44:39,646 INFO livekit.agents EOU metrics
2025-12-04 16:44:39,646 INFO basic-agent [STATE] agent_state changed: listening -> thinking | user_state=listening
2025-12-04 16:44:40,662 INFO livekit.agents LLM metrics
2025-12-04 16:44:41,836 INFO basic-agent [STATE] agent_state changed: thinking -> speaking | user_state=listening
2025-12-04 16:44:44,494 INFO livekit.agents STT metrics
2025-12-04 16:44:47,337 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=speaking
2025-12-04 16:44:47,492 INFO livekit.agents STT metrics
2025-12-04 16:44:48,446 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=speaking
2025-12-04 16:44:48,486 INFO basic-agent [STT] PARTIAL text='Tell me about rocks.' | agent_state=speaking | user_state=listening
2025-12-04 16:44:48,796 INFO livekit.agents STT metrics
2025-12-04 16:44:48,803 INFO basic-agent [STT] FINAL text='Tell me about rocks.' | agent_state=speaking | user_state=listening
2025-12-04 16:44:48,806 INFO basic-agent [LOGIC] words=['tell', 'me', 'about', 'rocks'] | is_soft=False | has_hard=False | agent_state=speaking
2025-12-04 16:44:48,806 INFO basic-agent [LOGIC] NON-SOFT utterance while speaking -> treating as interrupt (session.interrupt(force=True))
2025-12-04 16:44:48,820 INFO basic-agent [STATE] agent_state changed: speaking -> listening | user_state=listening
2025-12-04 16:44:48,936 WARNING livekit.agents preemptive generation enabled but chat context or tools have changed after `on_user_turn_completed`
2025-12-04 16:44:48,936 INFO livekit.agents EOU metrics
2025-12-04 16:44:48,950 INFO basic-agent [STATE] agent_state changed: listening -> thinking | user_state=listening
2025-12-04 16:44:50,238 INFO livekit.agents LLM metrics
2025-12-04 16:44:51,024 INFO basic-agent [STATE] agent_state changed: thinking -> speaking | user_state=listening
2025-12-04 16:44:53,776 INFO livekit.agents STT metrics
2025-12-04 16:44:54,816 INFO livekit.agents TTS metrics
2025-12-04 16:44:55,647 INFO basic-agent [STATE] user_state changed: listening -> speaking | agent_state=speaking
2025-12-04 16:44:56,887 INFO basic-agent [STT] PARTIAL text='Okay. Right. But wait.' | agent_state=speaking | user_state=speaking
2025-12-04 16:44:56,946 INFO basic-agent [STATE] user_state changed: speaking -> listening | agent_state=speaking
2025-12-04 16:44:57,093 INFO livekit.agents STT metrics
2025-12-04 16:44:57,096 INFO basic-agent [STT] FINAL text='Okay. Right. But wait.' | agent_state=speaking | user_state=listening
2025-12-04 16:44:57,096 INFO basic-agent [LOGIC] words=['okay', 'right', 'but', 'wait'] | is_soft=False | has_hard=True | agent_state=speaking
2025-12-04 16:44:57,096 INFO basic-agent [LOGIC] HARD interrupt while speaking -> calling session.interrupt(force=True)
2025-12-04 16:44:57,113 INFO basic-agent [STATE] agent_state changed: speaking -> listening | user_state=listening
2025-12-04 16:44:57,192 WARNING livekit.agents preemptive generation enabled but chat context or tools have changed after `on_user_turn_completed`
2025-12-04 16:44:57,194 INFO livekit.agents EOU metrics
2025-12-04 16:44:57,203 INFO basic-agent [STATE] agent_state changed: listening -> thinking | user_state=listening
2025-12-04 16:44:58,697 INFO livekit.agents LLM metrics
2025-12-04 16:44:59,296 INFO basic-agent [STATE] agent_state changed: thinking -> speaking | user_state=listening
2025-12-04 16:44:59,876 INFO livekit.agents TTS metrics
2025-12-04 16:45:02,092 INFO livekit.agents STT metrics
2025-12-04 16:45:02,846 INFO basic-agent [STATE] agent_state changed: speaking -> listening | user_state=listening
2025-12-04 16:45:05,086 INFO livekit.agents STT metrics
2025-12-04 16:45:08,216 INFO livekit.agents shutting down worker
2025-12-04 16:45:08,235 INFO basic-agent [METRICS] Usage: UsageSummary(llm_prompt_tokens=2032, llm_prompt_cached_tokens=0, llm_input_audio_tokens=0, llm_input_cached_audio_tokens=0, llm_input_text_tokens=0, llm_input_cached_text_tokens=0, llm_input_image_tokens=0, llm_input_cached_image_tokens=0, llm_completion_tokens=235, llm_output_audio_tokens=0, llm_output_image_tokens=0, llm_output_text_tokens=0, tts_characters_count=838, tts_audio_duration=52.919999999999945, stt_audio_duration=69.74999999999991)
2025-12-04 16:45:08,248 INFO livekit.agents process exiting
42 changes: 42 additions & 0 deletions examples/voice_agents/interrupt_config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,42 @@
# interrupt_config.py

"""
Configuration for interrupt / backchannel word lists.

Change SOFT_WORDS and HARD_WORDS here without touching the agent logic.
All words must be lowercase because transcripts are lowercased before matching.
"""

SOFT_WORDS: set[str] = {
"yeah",
"yea",
"yah",
"ya",
"ok",
"okay",
"k",
"kk",
"hmm",
"mm",
"mmm",
"uh",
"uhh",
"uhm",
"um",
"huh",
"right",
"yep",
"yup",
"uh-huh",
"uhhuh",
}

HARD_WORDS: set[str] = {
"stop",
"wait",
"no",
"nope",
"cancel",
"pause",
"hold",
}
Loading