Examples: context compression

Shrink what each model call sends. Back to the Examples gallery. Files here: 12.

File	Shows	Install	Run	Docs
examples/compression_agent_framework.py	Compress what Microsoft Agent Framework sends on every model call, with chat middleware.	pip install "agent-adapter[agent-framework]"	import only	Compression by framework
examples/compression_backend.py	No agent framework: a script and a FastAPI route call your model with compression on.	pip install agent-adapter; optional: pip install fastapi	python examples/compression_backend.py	Compress a long context
examples/compression_camel.py	Compress the CAMEL ChatAgent memory records it builds each model call's context from.	pip install "agent-adapter[camel]"	import only	Compression by framework
examples/compression_crewai.py	Compress a CrewAI agent's messages before each LLM call; they are OpenAI chat dicts.	pip install agent-adapter	import only	Compression by framework
examples/compression_dspy.py	Compress the chat messages DSPy builds for each LM call, with an adapter subclass.	pip install "agent-adapter[dspy]"	import only	Compression by framework
examples/compression_framework_hooks.py	Compress what Pydantic AI, Agno, Strands, smolagents, AutoGen and AG2 send to the model.	pip install agent-adapter; optional: pip install "agent-adapter[ag2,agno,autogen,pydantic-ai,smolagents,strands]"	python examples/compression_framework_hooks.py	Compression by framework
examples/compression_frameworks.py	Compress a framework's own message list where it calls the model; no framework installed.	pip install agent-adapter	python examples/compression_frameworks.py	Compression by framework
examples/compression_langchain.py	LangChain alone: compress the message list before model.invoke; no LangGraph needed.	pip install "agent-adapter[langchain]"	python examples/compression_langchain.py	Compression by framework
examples/compression_langgraph.py	LangGraph: compress state["messages"] in your agent node, before the model.	pip install "agent-adapter[langgraph]"	python examples/compression_langgraph.py	Compression by framework
examples/compression_llamaindex.py	Compress a LlamaIndex ChatMessage list before the LLM call; the list keeps its type.	pip install "agent-adapter[llamaindex]"	import only	Compression by framework
examples/compression_one_call.py	One model call, four policies: trim only, cut only, summarise only, and all combined.	pip install agent-adapter	python examples/compression_one_call.py	Compress a long context
examples/compression_semantic_kernel.py	Compress a Semantic Kernel ChatHistory copy before the chat service call.	pip install "agent-adapter[semantic-kernel]"	import only	Compression by framework