OKF integrations

Every adapter has the same entry point, okf_tools(bundle, prefix=""), which returns the framework’s own tool objects. The adapter imports its framework only when you call it. Install an extra with pip install "astromesh-okf[<extra>]".

Framework	Extra	Adapter import
LangChain / LangGraph	`langchain`	`from astromesh_okf.integrations.langchain import okf_tools`
CrewAI	`crewai`	`from astromesh_okf.integrations.crewai import okf_tools`
Google ADK	`adk`	`from astromesh_okf.integrations.adk import okf_tools`
OpenAI Agents SDK	`openai-agents`	`from astromesh_okf.integrations.openai_agents import okf_tools`
PydanticAI (Python 3.11+)	`pydantic-ai`	`from astromesh_okf.integrations.pydantic_ai import okf_tools`
LlamaIndex	`llamaindex`	`from astromesh_okf.integrations.llamaindex import okf_tools`
smolagents	`smolagents`	`from astromesh_okf.integrations.smolagents import okf_tools`
Microsoft Agent Framework	`agent-framework`	`from astromesh_okf.integrations.agent_framework import okf_tools`
MCP (Claude Desktop, Claude Code, Cursor, any client)	`mcp`	`astromesh-okf mcp ./my-bundle`

Anything that takes plain Python functions needs no extra: OKFTools("./my-bundle").functions() returns kb_index, kb_read and kb_find.

The snippets below assume this system prompt. Model packages are separate installs.

from astromesh_okf import Bundle

SYSTEM = "Answer using the knowledge base tools. Bundle index:\n" + Bundle("./my-bundle").index("/")

LangChain / LangGraph

from langchain.agents import create_agent
from astromesh_okf.integrations.langchain import okf_tools

agent = create_agent("openai:gpt-4.1-mini", okf_tools("./my-bundle"), system_prompt=SYSTEM)
result = agent.invoke({"messages": [("user", "Can I get a refund?")]})

CrewAI

from crewai import Agent
from astromesh_okf.integrations.crewai import okf_tools

agent = Agent(
    role="Support agent",
    goal="Answer from the bundle",
    backstory=SYSTEM,
    tools=okf_tools("./my-bundle"),
)

Google ADK

from google.adk.agents import Agent
from astromesh_okf.integrations.adk import okf_tools

agent = Agent(
    name="support", model="gemini-2.5-flash", instruction=SYSTEM, tools=okf_tools("./my-bundle")
)

OpenAI Agents SDK

from agents import Agent, Runner
from astromesh_okf.integrations.openai_agents import okf_tools

agent = Agent(name="support", instructions=SYSTEM, tools=okf_tools("./my-bundle"))
print(Runner.run_sync(agent, "Can I get a refund?").final_output)

PydanticAI

from pydantic_ai import Agent
from astromesh_okf.integrations.pydantic_ai import okf_tools

agent = Agent("openai:gpt-4.1-mini", tools=okf_tools("./my-bundle"), system_prompt=SYSTEM)
print(agent.run_sync("Can I get a refund?").output)

LlamaIndex

from llama_index.core.agent.workflow import FunctionAgent
from llama_index.llms.openai import OpenAI
from astromesh_okf.integrations.llamaindex import okf_tools

agent = FunctionAgent(
    tools=okf_tools("./my-bundle"), llm=OpenAI(model="gpt-4.1-mini"), system_prompt=SYSTEM
)


async def main():
    print(await agent.run("Can I get a refund?"))

smolagents

from smolagents import OpenAIModel, ToolCallingAgent
from astromesh_okf.integrations.smolagents import okf_tools

agent = ToolCallingAgent(
    tools=okf_tools("./my-bundle"), model=OpenAIModel("gpt-4.1-mini"), instructions=SYSTEM
)
print(agent.run("Can I get a refund?"))

Microsoft Agent Framework

from agent_framework import Agent
from agent_framework.openai import OpenAIChatCompletionClient
from astromesh_okf.integrations.agent_framework import okf_tools

agent = Agent(
    OpenAIChatCompletionClient(model="gpt-4.1-mini"), SYSTEM, tools=okf_tools("./my-bundle")
)

MCP

See the quickstart for Claude Desktop, Claude Code and Cursor. From Python, build_server returns a FastMCP server you can run or mount:

from astromesh_okf.mcp import build_server

build_server("./my-bundle", prefix="support_", name="support-kb").run()

Argument coercion differs by framework

Models invent arguments: limit="2", tags="finance" instead of a list, include_deprecated="false" as a string; strict-schema frameworks send null for every optional argument.

Pass-through (plain functions and OKFTools, LlamaIndex, smolagents, Google ADK): the tool layer coerces. Numbers are parsed (with a default when they are not), a string tag becomes a list, "true", "1" and "yes" mean true and any other string means false, and null means the default.
Schema-validating (LangChain, CrewAI, OpenAI Agents, Agent Framework, PydanticAI and MCP): the framework validates first. It parses "2" and "false" itself, rejects what does not fit (limit="many", tags="finance") and reports the error to the model so it can retry.

Either way the agent loop does not crash.