How run-llama/llama_index works
LlamaIndex OSS is a Python framework for grounding LLM applications in user data, with retrieval-augmented querying and agent orchestration as central capabilities. Its documented workflow is to connect and ingest data, structure it into indices or graphs, retrieve relevant context, and produce an answer; agents and chat engines provide higher-level application patterns. The query-engine, ingestion, storage, and tool edges are supported by sampled source; broader capability links reflect README descriptions and directory paths, not inspected implementations.
Main entry points: query_engine Query Engines · pipeline.py Ingestion Pipeline · indices Indices and Graphs · citation_query_engine.py Citation Query
Components 22
Data Ingestion
- Ingestion Pipeline
llama-index-core/llama_index/core/ingestion/pipeline.py - Document Parsers
llama-index-core/llama_index/core/node_parser - Metadata Extractors
llama-index-core/llama_index/core/extractors
Data Structures
- Documents and Nodes
llama-index-core/llama_index/core/schema.py - Indices and Graphs
llama-index-core/llama_index/core/indices - Embeddings
llama-index-core/llama_index/core/embeddings
Query and Retrieval
- Retrievers
llama-index-core/llama_index/core/base/base_retriever.py - Node Postprocessors
llama-index-core/llama_index/core/postprocessor - Query Engines
llama-index-core/llama_index/core/query_engine - Citation Query
llama-index-core/llama_index/core/query_engine/citation_query_engine.py - Response Synthesis
llama-index-core/llama_index/core/response_synthesizers
Agents and Apps
- Query Engine Tools
llama-index-core/llama_index/core/tools/query_engine.py - Agent Workflows
llama-index-core/llama_index/core/agent/workflow - Chat Engines
llama-index-core/llama_index/core/chat_engine - Evaluation
llama-index-core/llama_index/core/evaluation
Core Services
- Storage Context
llama-index-core/llama_index/core/storage/storage_context.py - Document and Vector Stores
llama-index-core/llama_index/core/storage - Callbacks and Instrumentation
llama-index-core/llama_index/core/callbacks
Other components
- Application Developer
- Data Sources
- LLM Providers
- Integrations
Connections 23
- Application Developer → Ingestion Pipeline: configures ingestion
- Data Sources → Ingestion Pipeline: provides data
- Ingestion Pipeline → Document Parsers: transforms documents
- Ingestion Pipeline → Metadata Extractors: applies metadata
- Ingestion Pipeline → Documents and Nodes: produces nodes
- Documents and Nodes → Indices and Graphs: structures data
- Indices and Graphs → Embeddings: uses embeddings
- Indices and Graphs → Storage Context: uses storage
- Storage Context → Document and Vector Stores: persists data
- Query Engines → Retrievers: retrieves context
- Query Engines → Node Postprocessors: processes nodes
- Query Engines → Response Synthesis: synthesizes response
- Citation Query → Retrievers: retrieves nodes
- Citation Query → Node Postprocessors: processes nodes
- Citation Query → Response Synthesis: synthesizes answer
- Query Engine Tools → Query Engines: runs query
- Agent Workflows → Query Engine Tools: dispatches tools
- Chat Engines → Query Engines: answers with context
- Query Engines → LLM Providers: generates response
- Evaluation → Query Engines: evaluates outputs
- Query Engines → Callbacks and Instrumentation: emits events
- Application Developer → Integrations: selects plugins
- Integrations → Indices and Graphs: extends framework