Hand-picked tutorials, guides and examples on AI & Machine Learning — explained simply.
Showing 40 articles
Large language model APIs all follow the same basic shape: you send a list of messages and get a message back. This guide uses the official Anthropic SDK to call Claude, but the ideas carry over to any provider. 1. Install the SDK and set your key # TypeScript...
An LLM only knows what was in its training data. Retrieval-augmented generation (RAG) lets it answer questions about your data, such as docs, tickets or policies, by finding the relevant passages first and putting them in the prompt. The pipeline Ingest: split...
"Agent" is used for many things, but the core idea is simple: an AI agent is an LLM that decides which tools to call, sees the results, and keeps going until the task is done . Your code runs the tools; the model chooses what to do next. The loop Send the user...
Before MCP, every AI application needed custom glue code for every integration: one plugin for GitHub in app A, a different one in app B. The Model Context Protocol (MCP) is an open standard, introduced by Anthropic, that defines one way for AI applications to...
Long answers can take many seconds to generate. Streaming shows text as it is produced, so users see progress immediately instead of staring at a spinner. Streaming on the server The SDK's messages.stream() helper gives you an async iterator of events plus a t...
An embedding is a list of numbers (a vector) that represents the meaning of a piece of text. Texts with similar meanings get vectors that point in similar directions, even if they share no words. from sentence_transformers import SentenceTransformer model = Se...
Tool use (also called function calling) lets a model request that your code run a function. The model never runs code itself; it returns a structured request and waits for your result. 1. Describe the tools tools = [ { "name": "search_products", "description":...
In this tutorial you'll build an MCP server that exposes a get_github_repo tool, and run it over stdio so desktop AI apps can launch it. 1. Set up the project mkdir github-mcp && cd github-mcp npm init -y npm install @modelcontextprotocol/sdk zod npm install -...
The system prompt is the brief you give the model before the conversation starts. Most poor LLM results come from a vague brief, not a weak model. 1. Give context, not just commands Explain who the users are and why the task matters. A model that understands t...
Chunking decides what the retriever can find. Chunks that are too small lose context; chunks that are too large mix several topics and dilute the embedding. 1. Fixed-size chunks with overlap The simplest approach: split every N tokens or characters and repeat...
The model chooses tools based only on their names, descriptions and schemas. A vague description is the most common reason an agent picks the wrong tool or invents arguments. Write descriptions like documentation for a new teammate { "name": "create_refund", "...
The official Python SDK includes FastMCP , a high-level API where type hints and docstrings become the tool schema. A working server takes only a few lines. 1. Create the project uv init notes-mcp && cd notes-mcp uv add "mcp[cli]" 2. Write the server # server....
Asking a model to "reply in JSON" works most of the time. Production code needs it to work every time. Structured outputs constrain the response to a JSON Schema, so the result always parses. TypeScript with Zod import Anthropic from '@anthropic-ai/sdk'; impor...
You don't always need a separate vector database. The pgvector extension adds a vector type, distance operators and approximate indexes to PostgreSQL, so embeddings live next to the rest of your data. Set up the table CREATE EXTENSION IF NOT EXISTS vector; CRE...
The manual agent loop is always the same: call the model, run the requested tools, send results back, repeat. The Anthropic SDKs include a tool runner (currently in beta) that does this for you, with type-checked tool inputs. TypeScript with Zod import Anthrop...
Most MCP tutorials stop at tools, but the protocol has three server primitives, each controlled by a different party: Tools - the model decides when to call them. Resources - the application (or user) decides what data to attach. Prompts - the user explicitly...
Many apps send the same large prefix on every request: a long system prompt, tool definitions, or a document the user is asking about. Prompt caching stores that prefix so later requests read it from cache at a fraction of the normal input price and with lower...
Vector search is great at meaning ("cancel my subscription" finds "how to end your plan") but weak at exact tokens: error codes, product SKUs, function names. Keyword search is the opposite. Hybrid search runs both and merges the results. Keyword search in Pos...
When a request needs several independent lookups ("compare the weather in Delhi, Mumbai and Bengaluru"), the model can return multiple tool_use blocks in a single response. Running them one after another wastes time. Execute in parallel const toolUses = respon...
Once you have an MCP server, connect it to a host app so the model can use it. Here's how to do that in two popular hosts. Claude Desktop Open Settings, then Developer, then Edit Config. This opens claude_desktop_config.json . Add your server under mcpServers...
LLM APIs bill by the token , a chunk of text that is often a word or part of a word. Input and output tokens are priced separately, and output tokens cost several times more than input tokens. Count tokens before you send Token counts depend on the model's tok...
Embedding search compares a query vector with chunk vectors that were computed separately. It is fast, but it misses nuance. A reranker reads the query and each candidate chunk together and scores how well they match. It is slower, so you use it on a short lis...
Reading data is low-risk. Sending emails, moving money or deleting records is not. A human-in-the-loop step pauses the agent and asks for approval before risky tools run. Classify your tools Read-only (search, get_*): run automatically. Reversible writes (draf...
Stdio servers run on the user's machine. A remote MCP server runs on your infrastructure and is reached over HTTP, which is how SaaS products expose their tools to AI apps. The current HTTP transport is Streamable HTTP : one endpoint that accepts JSON-RPC over...
LLM APIs are network services under heavy load, so failures are normal. The trick is knowing which errors to retry and which to surface to the user. Retryable vs non-retryable Retry: 429 rate limit, 5xx server errors, overloaded errors, timeouts and dropped co...
Users trust an answer more when they can click through to the source. Citations also make hallucinations easy to spot: an unsupported claim has nothing to point at. Option 1: Ask for chunk IDs Answer the question using only the sources below. After each senten...
An agent can only do what its tools allow. Good guardrails assume the model will sometimes make mistakes or be manipulated, and limit the damage when it does. 1. Least privilege Give each agent only the tools its job needs. Scope credentials: a read-only datab...
Host apps like Claude Desktop include an MCP client, but you can embed one in your own app too. That lets your product use any MCP server as a tool provider. Connect and list tools import { Client } from '@modelcontextprotocol/sdk/client/index.js'; import { St...
Few-shot prompting means showing the model a handful of input/output examples before the real input. It is the fastest way to teach a specific style or labelling scheme without fine-tuning. A classification example Label each commit message as feat, fix, docs...
Without evaluation, every change to chunking, embeddings or prompts is a guess. A small labelled dataset turns RAG tuning into engineering. 1. Build an evaluation set Collect 50-200 real questions from logs, support tickets or users. For each, record the ID(s)...
Every turn of an agent adds to the conversation: tool calls, large tool results, the model's replies. Long tasks can grow past the context window, and even before that, a bloated context makes the model slower, more expensive and less focused. 1. Keep tool res...
Debugging through a chat app is slow: you can't see the raw messages, and the model decides when to call your tool. The MCP Inspector is an official browser-based tool that connects to your server directly. Run it # TypeScript / Node server npx @modelcontextpr...
Reasoning models can "think" before answering: they work through a problem internally, then write the final response. Thinking improves results on hard tasks such as multi-step maths, debugging and planning, at the cost of extra tokens and time. Adaptive think...
Exact nearest-neighbour search checks every vector, which gets slow past a few hundred thousand rows. Approximate nearest neighbour (ANN) indexes trade a little recall for a huge speed-up. pgvector offers two: HNSW and IVFFlat . HNSW (Hierarchical Navigable Sm...
Agents are exciting, but they cost more, are slower and are harder to test than fixed code. Many features are better built as workflows : LLM calls arranged in a flow that your code controls. Start with the simplest thing One LLM call - classification, summari...
MCP connects models to real systems, so it inherits real security risks. Some apply to server authors, some to anyone installing servers. Risks for users and host apps Malicious servers. A local stdio server runs with your user's permissions. Only install serv...
Modern LLMs read images and PDFs as well as text. That makes it easy to build features like "extract the line items from this invoice" or "explain this error screenshot". Image from a URL const response = await client.messages.create({ model: 'claude-opus-5-5'...
A chunk like "Revenue grew 3% over the previous quarter" is useless on its own: which company, which quarter? Contextual retrieval , a technique published by Anthropic, fixes this by adding a short explanation of where the chunk sits in its document before emb...
Agents are non-deterministic, so traditional "input X gives exactly output Y" tests only go so far. A solid testing strategy combines normal tests for your code with task-level evaluations of the agent's behaviour. 1. Unit-test the tools Tools are ordinary fun...
If an MCP server is reachable over HTTP, the Claude API can connect to it for you. You list the server in the request and the API handles tool discovery and tool calls, so there's no MCP client or agent loop in your code. The request The MCP connector is a bet...
