Claude API Response: stop_reason, usage, and Content Blocks
September 30, 2026
Decode every Claude API response field: what each stop_reason value means, how to read usage token counters, and which content block types your code must handle.
Read →Deep-dives on CCA-F exam prep, study patterns that actually work, and behind-the-scenes posts on how Plinth's weak-spot drill engine works.
September 30, 2026
Decode every Claude API response field: what each stop_reason value means, how to read usage token counters, and which content block types your code must handle.
Read →September 25, 2026
The Claude Message Batches API runs up to 100,000 requests asynchronously at 50% of standard pricing — the right surface when throughput beats latency for CCA-F and production workloads.
Read →September 23, 2026
How the Claude streaming API delivers tokens over SSE, the six event types in sequence, when to stream vs. batch, and what the CCA-F exam tests.
Read →September 20, 2026
How to send images, PDFs, and documents to the Claude API: accepted formats, base64-vs-URL, document blocks, token costs, and CCA-F exam concepts.
Read →September 9, 2026
Claude's agentic safety model — minimal footprint, reversible-first actions, and pause-and-verify triggers — explained for the CCA-F exam.
Read →September 7, 2026
Claude's behaviors split into hardcoded limits no system prompt can override and softcoded defaults operators can adjust — here's the exact taxonomy the CCA-F exam tests.
Read →August 31, 2026
Learn Claude's per-model context window limits, what counts toward the token budget, how to count tokens via the API, and which management strategy fits your use case.
Read →August 26, 2026
Anthropic Claude prompt caching cuts read costs 90% but adds a 25% write premium — master the break-even math and decision framework for the CCA-F exam.
Read →August 24, 2026
Model context protocol registry explained: what MCP servers are, how server discovery works, and what the CCA-F exam expects you to know.
Read →August 21, 2026
How Claude's three-tier Anthropic–operator–user model works, what each party can permit or restrict, and what the CCA-F exam tests.
Read →August 19, 2026
Claude extended thinking gives the model a private reasoning chain that improves accuracy on hard tasks — here's the decision framework for enabling it without paying for it where it won't help.
Read →August 17, 2026
A complete guide to the CCA-F exam: registration steps, question format, scoring, and what the Anthropic certification credential delivers after you pass.
Read →August 14, 2026
Claude Code prompt caching stores your system prompt server-side with a 5-minute TTL, cutting input costs by up to 90% — what the CCA-F exam tests.
Read →August 12, 2026
Master the MCP architecture diagram — host, client, and server roles plus both transports — to tackle MCP questions on the Anthropic CCA-F exam.
Read →August 10, 2026
The official Model Context Protocol documentation mapped section-by-section to CCA-F exam domains — know exactly what to read, skim, and skip.
Read →August 7, 2026
A knowledge map of Claude Code's agent loop, tool permissions, and MCP integration for the Claude Certified Architect exam.
Read →August 5, 2026
The Anthropic Architect Certification (CCA-F) tests technical depth across seven domains — here's the skill checklist candidates need for MCP, multi-agent orchestration, and responsible AI design.
Read →August 5, 2026
A practical guide to selecting the right claude api model — Haiku, Sonnet, or Opus — covering capability tradeoffs, API credit costs, model IDs in code, and CCA-F exam essentials.
Read →July 29, 2026
A technical explainer mapping the three layers of the Claude Code API — Anthropic REST API, CLI invocation surface, and Agent SDK — with CCA-F exam guidance.
Read →July 27, 2026
The Claude Certified Architect exam splits across five weighted domains. Here is what each one covers and how to turn the weighting into a study plan.
Read →July 27, 2026
Prompt caching lets Claude reuse stable context across API requests — cutting input token costs by up to 90% when you structure your prompts correctly, using the cache_control parameter.
Read →July 25, 2026
A practical study guide for the Claude Certified Architect — Foundations exam: what it tests, how long to prepare, and which skills to build first.
Read →July 25, 2026
How the Claude Agent SDK structures agentic work: tool runner, Managed Agents, MCP integration, and the pause_turn behavior that trips up most builders.
Read →July 23, 2026
Claude tool use lets the model call your functions — define a tool with a JSON Schema, detect the tool_use stop reason, execute it, and loop until Claude is done.
Read →July 21, 2026
Model Context Protocol (MCP) lets Claude connect to external services like GitHub and Linear without embedding credentials in the agent definition. Here is how it works.
Read →July 21, 2026
What the Claude Certified Architect — Foundations exam tests, how hard it is, how long to study, and whether the CCA-F credential is worth pursuing.
Read →Plinth Prep is an independent study resource and is not affiliated with, endorsed by, or sponsored by Anthropic.