wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fixed MCPClient to reuse session across tool calls

Browse Source
This commit is contained in:
Om Chauhan
2026-03-15 17:57:51 +05:30
committed by vipyne
parent 57068f1b38
commit e22f9f84bb
5 changed files with 292 additions and 353 deletions
Download Patch File Download Diff File Expand all files Collapse all files

105
examples/mcp/mcp-multiple-mcp.py
View File

@@ -146,90 +146,77 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
),
)
try:
rijksmuseum_mcp = MCPClient(
async with (
MCPClient(
server_params=StdioServerParameters(
command=shutil.which("npx"),
# https://github.com/r-huijts/rijksmuseum-mcp
args=["-y", "mcp-server-rijksmuseum"],
env={"RIJKSMUSEUM_API_KEY": os.getenv("RIJKSMUSEUM_API_KEY")},
)
)
except Exception as e:
logger.error(f"error setting up rijksmuseum mcp")
logger.exception("error trace:")
try:
) as rijksmuseum_mcp,
# Github MCP docs: https://github.com/github/github-mcp-server
# Enable Github Copilot on your GitHub account. Free tier is ok. (https://github.com/settings/copilot)
# Generate a personal access token. It must be a Fine-grained token, classic tokens are not supported. (https://github.com/settings/personal-access-tokens)
# Set permissions you want to use (eg. "all repositories", "profile: read/write", etc)
github_mcp = MCPClient(
MCPClient(
server_params=StreamableHttpParameters(
url="https://api.githubcopilot.com/mcp/",
headers={
"Authorization": f"Bearer {os.getenv('GITHUB_PERSONAL_ACCESS_TOKEN')}"
},
)
)
except Exception as e:
logger.error(f"error setting up mcp.run")
logger.exception("error trace:")
rijksmuseum_tools = {}
github_tools = {}
try:
),
) as github_mcp,
):
rijksmuseum_tools = await rijksmuseum_mcp.register_tools(llm)
github_tools = await github_mcp.register_tools(llm)
except Exception as e:
logger.error(f"error registering tools")
logger.exception("error trace:")
all_standard_tools = rijksmuseum_tools.standard_tools + github_tools.standard_tools
all_tools = ToolsSchema(standard_tools=all_standard_tools)
all_standard_tools = rijksmuseum_tools.standard_tools + github_tools.standard_tools
all_tools = ToolsSchema(standard_tools=all_standard_tools)
context = LLMContext(tools=all_tools)
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
mcp_image_processor = UrlToImageProcessor(aiohttp_session=session)
context = LLMContext(tools=all_tools)
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
mcp_image_processor = UrlToImageProcessor(aiohttp_session=session)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User spoken responses
llm, # LLM
tts, # TTS
mcp_image_processor, # URL image -> output
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses and tool context
]
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User spoken responses
llm, # LLM
tts, # TTS
mcp_image_processor, # URL image -> output
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses and tool context
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
await runner.run(task)
async def bot(runner_args: RunnerArguments):

110
examples/mcp/mcp-stdio.py
View File

@@ -162,73 +162,63 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
),
)
try:
mcp = MCPClient(
server_params=StdioServerParameters(
command=shutil.which("npx"),
# https://github.com/r-huijts/rijksmuseum-mcp
args=["-y", "mcp-server-rijksmuseum"],
env={"RIJKSMUSEUM_API_KEY": os.getenv("RIJKSMUSEUM_API_KEY")},
),
# Optional
tools_filter=mcp_tools_filter, # Optional
tools_output_filters={"open_image_in_browser": open_image_output_filter},
)
except Exception as e:
logger.error(f"error setting up mcp")
logger.exception("error trace:")
mcp_image = UrlToImageProcessor(aiohttp_session=session)
tools = {}
try:
tools = await mcp.register_tools(llm)
except Exception as e:
logger.error(f"error registering tools")
logger.exception("error trace:")
context = LLMContext(tools=tools)
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User spoken responses
llm, # LLM
tts, # TTS
mcp_image, # URL image -> output
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses and tool context
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
async with MCPClient(
server_params=StdioServerParameters(
command=shutil.which("npx"),
# https://github.com/r-huijts/rijksmuseum-mcp
args=["-y", "mcp-server-rijksmuseum"],
env={"RIJKSMUSEUM_API_KEY": os.getenv("RIJKSMUSEUM_API_KEY")},
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
# Optional
tools_filter=mcp_tools_filter, # Optional
tools_output_filters={"open_image_in_browser": open_image_output_filter},
) as mcp:
tools = await mcp.register_tools(llm)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
await task.queue_frames([LLMRunFrame()])
context = LLMContext(tools=tools)
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User spoken responses
llm, # LLM
tts, # TTS
mcp_image, # URL image -> output
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses and tool context
]
)
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
await runner.run(task)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):

112
examples/mcp/mcp-streamable-http-gemini-live.py
View File

@@ -63,28 +63,6 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
),
)
try:
# Github MCP docs: https://github.com/github/github-mcp-server
# Enable Github Copilot on your GitHub account. Free tier is ok. (https://github.com/settings/copilot)
# Generate a personal access token. It must be a Fine-grained token, classic tokens are not supported. (https://github.com/settings/personal-access-tokens)
# Set permissions you want to use (eg. "all repositories", "profile: read/write", etc)
mcp = MCPClient(
server_params=StreamableHttpParameters(
url="https://api.githubcopilot.com/mcp/",
headers={"Authorization": f"Bearer {os.getenv('GITHUB_PERSONAL_ACCESS_TOKEN')}"},
)
)
except Exception as e:
logger.error(f"error setting up mcp")
logger.exception("error trace:")
tools = {}
try:
tools = await mcp.get_tools_schema()
except Exception as e:
logger.error(f"error registering tools")
logger.exception("error trace:")
system = f"""
You are a helpful LLM in a voice call.
Your goal is to answer questions about the user's GitHub repositories and account.
@@ -94,53 +72,65 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
Just respond with short sentences when you are carrying out tool calls.
"""
llm = GeminiLiveLLMService(
api_key=os.getenv("GOOGLE_API_KEY"),
system_instruction=system,
tools=tools,
)
# Github MCP docs: https://github.com/github/github-mcp-server
# Enable Github Copilot on your GitHub account. Free tier is ok. (https://github.com/settings/copilot)
# Generate a personal access token. It must be a Fine-grained token, classic tokens are not supported. (https://github.com/settings/personal-access-tokens)
# Set permissions you want to use (eg. "all repositories", "profile: read/write", etc)
async with MCPClient(
server_params=StreamableHttpParameters(
url="https://api.githubcopilot.com/mcp/",
headers={"Authorization": f"Bearer {os.getenv('GITHUB_PERSONAL_ACCESS_TOKEN')}"},
)
) as mcp:
tools = await mcp.get_tools_schema()
await mcp.register_tools_schema(tools, llm)
llm = GeminiLiveLLMService(
api_key=os.getenv("GOOGLE_API_KEY"),
system_instruction=system,
tools=tools,
)
context = LLMContext([{"role": "developer", "content": "Please introduce yourself."}])
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
await mcp.register_tools_schema(tools, llm)
pipeline = Pipeline(
[
transport.input(), # Transport user input
user_aggregator, # User spoken responses
llm, # LLM
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses and tool context
]
)
context = LLMContext([{"role": "user", "content": "Please introduce yourself."}])
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
user_aggregator, # User spoken responses
llm, # LLM
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses and tool context
]
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
await task.queue_frames([LLMRunFrame()])
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
await task.queue_frames([LLMRunFrame()])
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
await runner.run(task)
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):

98
examples/mcp/mcp-streamable-http.py
View File

@@ -77,69 +77,59 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
system_instruction=system_prompt,
)
try:
# Github MCP docs: https://github.com/github/github-mcp-server
# Enable Github Copilot on your GitHub account. Free tier is ok. (https://github.com/settings/copilot)
# Generate a personal access token. It must be a Fine-grained token, classic tokens are not supported. (https://github.com/settings/personal-access-tokens)
# Set permissions you want to use (eg. "all repositories", "profile: read/write", etc)
mcp = MCPClient(
server_params=StreamableHttpParameters(
url="https://api.githubcopilot.com/mcp/",
headers={"Authorization": f"Bearer {os.getenv('GITHUB_PERSONAL_ACCESS_TOKEN')}"},
)
# Github MCP docs: https://github.com/github/github-mcp-server
# Enable Github Copilot on your GitHub account. Free tier is ok. (https://github.com/settings/copilot)
# Generate a personal access token. It must be a Fine-grained token, classic tokens are not supported. (https://github.com/settings/personal-access-tokens)
# Set permissions you want to use (eg. "all repositories", "profile: read/write", etc)
async with MCPClient(
server_params=StreamableHttpParameters(
url="https://api.githubcopilot.com/mcp/",
headers={"Authorization": f"Bearer {os.getenv('GITHUB_PERSONAL_ACCESS_TOKEN')}"},
)
except Exception as e:
logger.error(f"error setting up mcp")
logger.exception("error trace:")
tools = {}
try:
) as mcp:
tools = await mcp.register_tools(llm)
except Exception as e:
logger.error(f"error registering tools")
logger.exception("error trace:")
context = LLMContext(tools=tools)
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
context = LLMContext(tools=tools)
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User spoken responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses and tool context
]
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User spoken responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses and tool context
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
await runner.run(task)
async def bot(runner_args: RunnerArguments):