Workflow
LlamaIndex
icon
Search documents
LlamaIndex Live Stream
LlamaIndex· 2025-07-21 15:05
由于提供的 Content 为空,无法进行总结和分析,因此无法提供符合要求的输出。
Running the Open-Source LlamaCloud MCP Server
LlamaIndex· 2025-07-08 11:28
Llama Cloud MCP Server Overview - Llama Cloud offers an open-sourced MCP server in Python, enabling the use of Llama Cloud indexes and extraction agents as tools within any MCP client [1] - The MCP server allows users to integrate Llama Cloud's indexing and extraction capabilities into various applications [1] Key Components and Configuration - The system utilizes indexes such as a Google Drive index connected to Llama Index workflows and a filings index containing SEC filings [2] - Extraction agents, including an invoice extractor and a CV extractor, can be added and utilized as MCP tools [3][8][9] - Configuration involves defining indexes with names and descriptions, which are crucial for the LLM to determine the appropriate MCP tool to use [4] - A file system tool is used, granting access to specific folders on the user's machine, enabling interaction with local files [5] Functionality and Use Cases - The system can be used to extract information from files, such as identifying the invoiced party from an invoice [7][8] - The platform supports complex tasks involving multiple extractions, such as extracting information from both an invoice and a CV [11] - Users can add multiple extraction agents and indexes to the MCP server [14] Future Enhancements - Future development could include the ability to write new files and data into existing indexes using the file systems tool [15]
Observability in Agentic Applications with LlamaIndex and OpenTelemetry
LlamaIndex· 2025-06-30 13:40
Agent Capabilities - Lama Index introduces a syllabus extraction agent designed to summarize university course syllabi [1] - The agent utilizes a syllabus extractor tool from Lama Cloud to extract information [2] - The agent can answer questions about the syllabus after ingesting the file [6][7] Observability and Tracing - The agent provides observability results through traces stored in a SQL database [3] - Users can query the database to view agent traces and filter by duration (e g, less than 300 seconds) [3][4] - The tracing is based on an open-source integration with Open Telemetry [5] - Events are piped into a PostgreSQL database [5] Technology Stack - Llama Index uses Llama Cloud for syllabus extraction [2] - The tracing system integrates with Open Telemetry [5] - PostgreSQL is used as the database for storing traces [5]