Tutor Agent using Pipecat

1
import os
2
from dotenv import load_dotenv
3
from loguru import logger
4
from pipecat.frames.frames import LLMRunFrame
5
from pipecat.pipeline.pipeline import Pipeline
6
from pipecat.pipeline.runner import PipelineRunner
7
from pipecat.pipeline.task import PipelineTask
8
from pipecat.processors.aggregators.llm_context import LLMContext
9
from pipecat.processors.aggregators.llm_response_universal import (
10
    LLMContextAggregatorPair,
11
)
12
from pipecat.runner.types import RunnerArguments
13
from pipecat.runner.utils import create_transport
14
from pipecat.services.sarvam.stt import SarvamSTTService
15
from pipecat.services.sarvam.tts import SarvamTTSService
16
from pipecat.services.openai.llm import OpenAILLMService
17
from pipecat.transports.base_transport import TransportParams
18
from pipecat.transports.daily.transport import DailyParams
19
20
load_dotenv(override=True)
21
22
async def bot(runner_args: RunnerArguments):
23
    """Main bot entry point."""
24
    
25
    # Create transport (supports both Daily and WebRTC)
26
    transport = await create_transport(
27
        runner_args,
28
        {
29
            "daily": lambda: DailyParams(audio_in_enabled=True, audio_out_enabled=True),
30
            "webrtc": lambda: TransportParams(
31
                audio_in_enabled=True, audio_out_enabled=True
32
            ),
33
        },
34
    )
35
36
    # Initialize AI services
37
    stt = SarvamSTTService(
38
        api_key=os.getenv("SARVAM_API_KEY"),
39
        language="unknown",  # Auto-detect for multilingual students
40
        model="saaras:v3",
41
        mode="transcribe"
42
    )
43
    
44
    tts = SarvamTTSService(
45
        api_key=os.getenv("SARVAM_API_KEY"),
46
        target_language_code="en-IN",
47
        model="bulbul:v3",
48
        speaker="ishita"  # Clear and articulate voice for teaching
49
    )
50
    
51
    llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"), model="gpt-4o")
52
53
    # Set up conversation context with tutor personality
54
    messages = [
55
        {
56
            "role": "system",
57
            "content": """You are an expert tutor designed to help students understand and excel in their studies.
58
59
Your teaching expertise covers multiple subjects:
60
61
**Mathematics:**
62
- Arithmetic, Algebra, Geometry, Trigonometry
63
- Calculus, Statistics, Probability
64
- Problem-solving techniques
65
66
**Science:**
67
- Physics: Mechanics, Electricity, Optics, Thermodynamics
68
- Chemistry: Elements, Reactions, Organic Chemistry
69
- Biology: Cell Biology, Human Anatomy, Ecology
70
71
**Languages:**
72
- English Grammar and Composition
73
- Hindi Grammar and Literature
74
- Reading Comprehension
75
76
**Social Studies:**
77
- History, Geography, Civics
78
- Economics basics
79
80
Teaching approach:
81
- Start with the basics and build up to complex concepts
82
- Use real-world examples and analogies to explain abstract concepts
83
- Break down complex problems into smaller, manageable steps
84
- Encourage students and praise their efforts
85
- Ask questions to check understanding
86
- Adapt your explanations based on the student's level
87
- Use simple language and avoid overwhelming with jargon
88
- When solving numerical problems, show each step clearly
89
90
Communication style:
91
- Be patient, encouraging, and supportive
92
- Speak clearly and at a moderate pace
93
- Celebrate small victories and correct mistakes gently
94
- If a student is struggling, try a different explanation approach
95
- Make learning interesting by connecting it to everyday life
96
97
Start by greeting the student warmly and asking what subject or topic they'd like to learn or what problem they need help with.""",
98
        },
99
    ]
100
    context = LLMContext(messages)
101
    context_aggregator = LLMContextAggregatorPair(context)
102
103
    # Build pipeline
104
    pipeline = Pipeline(
105
        [
106
            transport.input(),
107
            stt,
108
            context_aggregator.user(),
109
            llm,
110
            tts,
111
            transport.output(),
112
            context_aggregator.assistant(),
113
        ]
114
    )
115
116
    task = PipelineTask(pipeline)
117
118
    @transport.event_handler("on_client_connected")
119
    async def on_client_connected(transport, client):
120
        logger.info("Student connected")
121
        messages.append(
122
            {"role": "system", "content": "Greet the student warmly and ask what subject or topic they'd like to learn today."}
123
        )
124
        await task.queue_frames([LLMRunFrame()])
125
126
    @transport.event_handler("on_client_disconnected")
127
    async def on_client_disconnected(transport, client):
128
        logger.info("Student disconnected")
129
        await task.cancel()
130
131
    runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
132
    await runner.run(task)
133
134
if __name__ == "__main__":
135
    from pipecat.runner.run import main
136
    main()

1
stt = SarvamSTTService(
2
    api_key=os.getenv("SARVAM_API_KEY"),
3
    language="hi-IN",  # Hindi
4
    model="saaras:v3",
5
    mode="transcribe"
6
)
7
8
tts = SarvamTTSService(
9
    api_key=os.getenv("SARVAM_API_KEY"),
10
    target_language_code="hi-IN",
11
    model="bulbul:v3",
12
    speaker="simran"  # Warm and friendly teacher voice
13
)
14
15
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"), model="gpt-4o")

1
stt = SarvamSTTService(
2
    api_key=os.getenv("SARVAM_API_KEY"),
3
    language="ta-IN",
4
    model="saaras:v3",
5
    mode="transcribe"
6
)
7
8
tts = SarvamTTSService(
9
    api_key=os.getenv("SARVAM_API_KEY"),
10
    target_language_code="ta-IN",
11
    model="bulbul:v3",
12
    speaker="ishita"
13
)
14
15
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"), model="gpt-4o")

1
stt = SarvamSTTService(
2
    api_key=os.getenv("SARVAM_API_KEY"),
3
    language="unknown",  # Auto-detects language
4
    model="saaras:v3",
5
    mode="transcribe"
6
)
7
8
tts = SarvamTTSService(
9
    api_key=os.getenv("SARVAM_API_KEY"),
10
    target_language_code="en-IN",
11
    model="bulbul:v3",
12
    speaker="ishita"
13
)
14
15
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"), model="gpt-4o")

1
# Student speaks Hindi/Tamil/etc. → Saaras converts to English → LLM processes
2
3
stt = SarvamSTTService(
4
    api_key=os.getenv("SARVAM_API_KEY"),
5
    model="saaras:v3",  # Speech-to-English translation
6
    mode="translate"
7
)
8
9
tts = SarvamTTSService(
10
    api_key=os.getenv("SARVAM_API_KEY"),
11
    target_language_code="en-IN",
12
    model="bulbul:v3",
13
    speaker="ishita"
14
)
15
16
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"), model="gpt-4o")

1
tts = SarvamTTSService(
2
    api_key=os.getenv("SARVAM_API_KEY"),
3
    target_language_code="en-IN",
4
    model="bulbul:v3",
5
    speaker="ishita",
6
    pace=0.9,            # Slightly slower for better understanding
7
    speech_sample_rate=24000  # 8000, 16000, 22050, 24000 Hz (default). v3 REST API also supports 32000, 44100, 48000 Hz
8
)