9.6.1 Frameworks Roadmap: Choose Only When Needed

Frameworks do not make an Agent smarter. They organize state, tools, workflows, memory, logs, and collaboration once the task has enough complexity to justify the abstraction.

See the Selection Map First

If a task has three fixed steps, plain Python functions may be better. Add a framework when state, branching, recovery, data connection, or role collaboration becomes hard to manage.

Run a Framework Route Check

Use this check before choosing a framework because it is popular.

task = {
    "needs_state": True,
    "needs_rag": False,
    "needs_roles": False,
    "needs_resume": True,
}

if task["needs_state"] or task["needs_resume"]:
    route = "LangGraph-style state graph"
elif task["needs_rag"]:
    route = "LlamaIndex-style data app"
elif task["needs_roles"]:
    route = "CrewAI or AutoGen-style collaboration"
else:
    route = "plain functions first"

print("route:", route)
print("reason:", "choose the smallest abstraction that exposes state")

Expected output:

route: LangGraph-style state graph
reason: choose the smallest abstraction that exposes state

Framework choice should be written into the README as a trade-off, not hidden inside dependencies.

Learn in This Order

Step	Read	Practice Output
1	Framework overview	Explain what a framework abstracts
2	LangChain / LangGraph	Model state, nodes, edges, branches, recovery
3	LlamaIndex	Connect documents, indexes, retrieval, evaluation
4	CrewAI / AutoGen	Compare role collaboration and multi-Agent conversation
5	Framework selection	Write a decision table and a no-framework baseline

Evidence to Keep

Keep this page’s proof of learning as a small evidence card:

Problem Shape

workflow graph, retrieval app, role team, or experiment

Framework Choice

what abstraction it adds and what control it hides

Trace

state, node, tool call, message, or run id

Failure Check

framework magic hides state, retries, or permissions

Decision

choose framework only after the single-agent loop is clear

Pass Check

You pass this chapter when you can implement the same small task with plain functions and with one framework, then explain which version is easier to debug and why.

Check reasoning and explanation

1. A passing answer describes the agent loop: goal, plan, tool call, observation, memory or state update, and stop condition.
2. The evidence should include a trace that another developer can inspect, not only the final answer.
3. A good self-check names one safety or reliability control such as tool schemas, permission boundaries, retries, evaluation cases, or a human-review point.