RAG vs MCP: RAG vs MCP: A Complete Guide for Developers in 2026

The following notes reconstruct a practical path around “RAG vs MCP: A Complete Guide for Developers in 2026”. Emphasis stays on contracts, checks, and drop-in code placeholders rather than motivational framing.

TL;DR

When working through TL;DR, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface.

Part 1: What RAG actually is

When working through Part 1: What RAG actually is, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface.

The analogy that makes it click

When working through The analogy that makes it click, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface.

Why it needs a vector database

When working through Why it needs a vector database, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface.

RAG in code

When working through RAG in code, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface. When working through RAG in code, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

from openai import OpenAI
client = OpenAI()
# Your knowledge base, already chunked.
DOCUMENTS = [
    "The P/E ratio divides share price by earnings per share. "
    "When earnings are negative, P/E is undefined and usually shown as N/A.",
    "EV/EBITDA is often preferred over P/E for capital-intensive companies "
    "because it is unaffected by capital structure and depreciation policy.",
    "The PEG ratio adjusts P/E by the expected earnings growth rate. "
    "A PEG below 1.0 is traditionally read as undervalued.",
]

def embed(text: str) -> list[float]:
    """Turn text into a vector."""
    response = client.embeddings.create(
        model="text-embedding-3-small",
        input=text,
    )
    return response.data[0].embedding

def cosine_similarity(a: list[float], b: list[float]) -> float:
    """How close are two vectors? 1.0 means identical direction."""
    dot = sum(x * y for x, y in zip(a, b))
    norm_a = sum(x * x for x in a) ** 0.5
    norm_b = sum(y * y for y in b) ** 0.5
    return dot / (norm_a * norm_b)

# Index once, reuse many times. In production this lives in a vector DB.
INDEX = [(doc, embed(doc)) for doc in DOCUMENTS]

def retrieve(question: str, k: int = 2) -> list[str]:
    """Step 1: find the most relevant chunks."""
    q_vector = embed(question)
    scored = [
        (cosine_similarity(q_vector, vector), doc)
        for doc, vector in INDEX
    ]
    scored.sort(reverse=True)
    return [doc for _, doc in scored[:k]]

def answer(question: str) -> str:
    """Steps 2 and 3: augment the prompt, then generate."""
    context = "\n\n".join(retrieve(question))
    prompt = (
        f"Answer using only the context below.\n\n"
        f"Context:\n{context}\n\n"
        f"Question: {question}"
    )
    response = client.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": prompt}],
    )
    return response.choices[0].message.content

print(answer("What do I use when a company has negative earnings?"))

Part 2: What MCP actually is

Part 2: What MCP actually is works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Separate chunking policy from retrieval policy. Changing one should not force a rewrite of the other when quality metrics move.

The analogy

The analogy works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Separate chunking policy from retrieval policy. Changing one should not force a rewrite of the other when quality metrics move.

The three things MCP servers expose

The three things MCP servers expose works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Separate chunking policy from retrieval policy. Changing one should not force a rewrite of the other when quality metrics move. The three things MCP servers expose works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

MCP in code

For MCP in code, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Cite the passages that actually grounded the answer. Without citations, operators cannot tell hallucination from an indexing gap.

from mcp.server.fastmcp import FastMCP
import httpx
mcp = FastMCP("finance-tools")

@mcp.tool()
def get_current_price(ticker: str) -> dict:
    """Get the latest price for a stock ticker.
    The docstring matters more than you'd think. It is what the
    model reads to decide whether to call this tool at all.
    """
    response = httpx.get(f"https://api.example.com/quote/{ticker}")
    return response.json()

@mcp.tool()
def compare_tickers(ticker_a: str, ticker_b: str) -> dict:
    """Compare two tickers on price, market cap and P/E ratio."""
    return {
        "a": get_current_price(ticker_a),
        "b": get_current_price(ticker_b),
    }

if __name__ == "__main__":
    mcp.run()

claude mcp add finance -- python /path/to/server.py

Part 3: The actual difference

For Part 3: The actual difference, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Cite the passages that actually grounded the answer. Without citations, operators cannot tell hallucination from an indexing gap.

The one-sentence rule

For The one-sentence rule, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Cite the passages that actually grounded the answer. Without citations, operators cannot tell hallucination from an indexing gap. For The one-sentence rule, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

Why “is RAG and MCP the same?” keeps getting asked

When working through Why “is RAG and MCP the same?” keeps getting asked, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface.

Part 4: The same assistant, built twice

When working through Part 4: The same assistant, built twice, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface.

Version A: the RAG approach

When working through Version A: the RAG approach, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface. When working through Version A: the RAG approach, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

# Building a knowledge base of financial concepts
CORPUS = [
    "Market capitalization equals share price multiplied by shares outstanding.",
    "The P/E ratio compares share price to earnings per share.",
    "Free cash flow is operating cash flow minus capital expenditures.",
    "A dividend yield above 6% often signals either a falling share price "
    "or an unsustainable payout ratio.",
    # ...plus a few thousand more chunks
]

# Index them, then:
answer("Explain what a high dividend yield might indicate")

answer("What is Apple's current dividend yield?")

Version B: the MCP approach

Version B: the MCP approach works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Separate chunking policy from retrieval policy. Changing one should not force a rewrite of the other when quality metrics move.

# EODHD publishes two endpoints. v2 uses OAuth, v1 uses an API key.
claude mcp add --transport http eodhd https://mcp.eodhd.com/v2/mcp

Version C: both, which is what you actually ship

Version C: both, which is what you actually ship works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Separate chunking policy from retrieval policy. Changing one should not force a rewrite of the other when quality metrics move.

def route(question: str) -> str:
    """Decide which subsystem answers this question."""
# Signals that the question is about live state
    live_signals = ["current", "today", "now", "latest", "price", "quote"]
    if any(signal in question.lower() for signal in live_signals):
        return "mcp"      # fetch it
    return "rag"          # look it up

Part 5: Related comparisons people conflate

Part 5: Related comparisons people conflate works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Separate chunking policy from retrieval policy. Changing one should not force a rewrite of the other when quality metrics move. Part 5: Related comparisons people conflate works best when treated as a measurable surface. Capture one golden transcript, one failure case, and the rollback note before expanding scope. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

MCP vs API

For MCP vs API, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Cite the passages that actually grounded the answer. Without citations, operators cannot tell hallucination from an indexing gap.

MCP vs agent

For MCP vs agent, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Cite the passages that actually grounded the answer. Without citations, operators cannot tell hallucination from an indexing gap.

RAG vs fine-tuning

For RAG vs fine-tuning, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Treat this stage as a contract between inputs and validated outputs. Name the artifacts, define success checks, and refuse silent partial completion. Cite the passages that actually grounded the answer. Without citations, operators cannot tell hallucination from an indexing gap. For RAG vs fine-tuning, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state. Keep configuration outside application code. Environment files, secret stores, and feature flags belong in one place operators can audit without reading the whole graph.

Part 6: Choosing your stack

When working through Part 6: Choosing your stack, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Document the happy path and the recovery path together. Retries, human gates, and dead-letter handling are part of the product, not later polish. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface.

FAQs

When working through FAQs, write down the contract first: required inputs, success signal, and what happens on partial failure. That checklist keeps later code changes honest. Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline. Measure recall on a fixed question set before tuning prompts. Prompt churn rarely fixes a weak retrieval surface.

Operational checklist

For Operational checklist, define the inputs, the owner of the step, and the exit criteria before changing code. Operators should be able to re-run the step from a known checkpoint without guessing hidden state.

Record timings and token or query cost next to functional results. Cost visibility early prevents surprise bills when the path moves from demo to shared environments.

Cite the passages that actually grounded the answer. Without citations, operators cannot tell hallucination from an indexing gap.

Log tool name, args hash, latency, and outcome for every call. Debugging agent loops without that trail wastes hours.

Pin dependency versions and record the image digest that ran the demo. Reproducibility beats tribal knowledge.

Prefer small, testable units over sprawling scripts. When a step fails, the failure should point at a single responsibility rather than a tangled pipeline.

Before promoting the stack, freeze versions, capture a golden transcript for the critical path, and confirm rollback steps. Shared environments need rate limits, tenancy checks, and a clear owner for secret rotation. Prefer boring reliability over clever one-off demos.

Batch note for 03e5c3844f8b: keep provider keys out of the repo, set a per-session token ceiling, and store transcripts next to the eval fixtures so later model swaps stay comparable.