Integrations

mycontext-ai is framework-agnostic. Every Context object can export itself into the native format of any major AI framework — no adapters, no glue code, no fighting with APIs.

How Integration Works

The Context object is the universal unit. Once you build one (from scratch or via a cognitive pattern), it can export itself to any framework:

from mycontext.templates.free.reasoning import RootCauseAnalyzer

# Build once
ctx = RootCauseAnalyzer().build_context(
    problem="API latency tripled after deployment",
    depth="comprehensive",
)

# Export anywhere
messages   = ctx.to_messages()          # OpenAI / universal messages
lc_format  = ctx.to_langchain()         # LangChain
li_format  = ctx.to_llamaindex()        # LlamaIndex
crew_cfg   = ctx.to_crewai()            # CrewAI Agent + Task
autogen_cfg = ctx.to_autogen()          # AutoGen AssistantAgent

Supported Frameworks

Framework	Install	Helper class	Native export
LangChain	pip install langchain-core	LangChainHelper	ctx.to_langchain()
LlamaIndex	pip install llama-index	LlamaIndexHelper	ctx.to_llamaindex()
CrewAI	pip install crewai	CrewAIHelper	ctx.to_crewai()
AutoGen	pip install pyautogen	AutoGenHelper	ctx.to_autogen()
DSPy	pip install dspy-ai	DSPyHelper	ctx.assemble()
Semantic Kernel	pip install semantic-kernel	SemanticKernelHelper	ctx.assemble()
Google ADK	pip install google-adk	GoogleADKHelper	ctx.assemble()

Two Integration Paths

Path 1: Native Context export methods

Every Context has built-in export methods:

from mycontext import Context
from mycontext.foundation import Guidance, Directive

ctx = Context(
    guidance=Guidance(role="Senior data analyst"),
    directive=Directive(content="Identify anomalies in the sales data"),
)

# Universal messages format
messages = ctx.to_messages()
# → [{"role": "system", "content": "..."}]

# LangChain
lc = ctx.to_langchain()
# → {"system_message": "...", "context": {...}}

# CrewAI
crew = ctx.to_crewai()
# → {"role": "...", "goal": "...", "backstory": "...", "expected_output": "..."}

# AutoGen
ag = ctx.to_autogen()
# → {"system_message": "...", "description": "...", ...}

# LlamaIndex
li = ctx.to_llamaindex()
# → {"template": "...", "system_prompt": "...", "query_instruction": "..."}

# Provider-specific
openai_cfg  = ctx.to_openai()     # {"messages": [...], "temperature": 0.7}
anthropic_cfg = ctx.to_anthropic() # {"system": "...", "messages": [...]}

Path 2: Helper classes

The helper classes add framework-specific functionality on top of the export methods:

from mycontext.integrations import (
    LangChainHelper,
    LlamaIndexHelper,
    CrewAIHelper,
    AutoGenHelper,
    DSPyHelper,
    SemanticKernelHelper,
    GoogleADKHelper,
)

auto_integrate() — One-liner

from mycontext.integrations import auto_integrate

ctx = RootCauseAnalyzer().build_context(problem="API latency spiked")

# Detect and export automatically
messages = auto_integrate(ctx, "langchain")
agent    = auto_integrate(ctx, "crewai", name="analyst", tools=my_tools)
prompt   = auto_integrate(ctx, "dspy")
adk_agent = auto_integrate(ctx, "google_adk", name="diagnostician")

Supported framework strings: "langchain", "llamaindex", "crewai", "autogen", "dspy", "semantic_kernel", "google_adk"

The Shared Pattern

Every integration follows the same pattern:

1. Build a Context (from scratch or via a cognitive pattern)
2. Export it to the framework's native format
3. Use the framework normally — mycontext just powered the prompt

This means your framework code doesn't change. You're just replacing ad-hoc system prompts with research-backed cognitive frameworks.

All Context Export Methods

Method	Returns	Use with
ctx.to_messages()	list[dict]	Any OpenAI-compatible API
ctx.to_openai()	dict	OpenAI Python client
ctx.to_anthropic()	dict	Anthropic Python client
ctx.to_langchain()	dict	LangChain / LangGraph
ctx.to_llamaindex()	dict	LlamaIndex query/chat engines
ctx.to_crewai()	dict	CrewAI Agent + Task
ctx.to_autogen()	dict	AutoGen AssistantAgent
ctx.assemble()	str	Any framework accepting a string
ctx.to_markdown()	str	Documentation, debugging
ctx.to_json()	str	Storage, APIs
ctx.to_yaml()	str	Config files
ctx.to_xml()	str	XML-based systems

Install Integration Dependencies

# Install only what you need
pip install mycontext-ai                  # Core only

pip install mycontext-ai langchain-core   # + LangChain
pip install mycontext-ai llama-index      # + LlamaIndex
pip install mycontext-ai crewai           # + CrewAI
pip install mycontext-ai pyautogen        # + AutoGen
pip install mycontext-ai dspy-ai          # + DSPy
pip install mycontext-ai semantic-kernel  # + Semantic Kernel
pip install mycontext-ai google-adk       # + Google ADK

# Or all at once
pip install "mycontext-ai[all]"