How I Cut My Claude API Bill in Half With Prompt Caching
Done right, Anthropic's prompt caching can roughly halve your monthly API spend on workloads with long, repeated system prompts. Here is the design playbook I use after six months in production, including the arithmetic that links hit rate to the bill and the break-even points (about 22% on the 5-minute TTL, 53% on the 1-hour TTL).
Running the Claude API in Python Production — Rate Limits, Retries, and Timeouts
If you put Claude API into a real backend service, how you handle 429, 503, and read timeouts decides your reliability ceiling. This is the design I settled on after operating it in production.
High-Availability Patterns for the Claude API — Making Sonnet/Haiku/Opus Fallback Work in Production
A single-model Claude API integration will fall over the first time rate limits or a regional hiccup land at peak hours. This is the production pattern for a Sonnet → Opus → Haiku fallback chain, with circuit breakers, streaming coverage, and the pitfalls you only learn the hard way.
Running Claude API Parallel Tool Use in Production — Controlling Concurrency, Designing for Partial Failure, and Cutting Latency
Claude API's parallel tool use can halve perceived latency, but partial failures and duplicate writes arrive just as fast. Concurrency control with asyncio.gather and a Semaphore, three partial-failure policies, idempotency, observability, and a way to measure your real parallel degree.
Intelligent Model Routing with Claude API — Auto-Selecting Sonnet 4.6 and Haiku 4.5 for Optimal Cost and Quality
Build an intelligent routing layer that automatically selects between Claude Sonnet 4.6 and Haiku 4.5 based on request complexity. Covers classifier design, circuit breakers, fallback chains, and cost monitoring for production deployments.
Building LLM Evaluation Pipelines with Claude API — Claude-as-Judge, Prompt A/B Testing, and Quality Scoring Patterns
Designing and implementing LLM evaluation pipelines on the Claude API — Claude-as-Judge, prompt A/B testing, quality scoring, and regression testing for production applications.
Precision Lives in the Second Stage: Reranking a Personal Knowledge Search with Claude
Embedding search alone leaves 'semantically close but not the answer' passages at the top. This is a two-stage design that gathers candidates broadly, then lets Claude reorder them by answerability, with structured scoring, abstention, and a cost estimate.
Coalescing Concurrent Claude API Calls: Single-Flight Against Duplicate Inference and Cache Stampede
A design for collapsing identical prompts that fire at the same instant into a single upstream Claude call, using single-flight (request coalescing). In-process and distributed implementations, jittered retries, and negative caching, with measured results.
Same Output, Different Path — Guarding Agent Trajectories with Invariants
When the default model changes, your final output can stay correct while the path your agent takes quietly shifts. Here is a trajectory regression harness built on recorded tool traces and deterministic invariants, with working code and measured numbers.
The 200K-Token Cliff That Doubled My Nightly Bill — A count_tokens Preflight to Avoid Long-Context Pricing
Sonnet 5's native 1M context switches the entire request to long-context pricing the moment input crosses 200K tokens. Here's how I caught that silent cost cliff before sending, using the billing-exempt count_tokens API, with working Python and TypeScript code.
claude-opus-4-7 Fast Mode Retires on 7/24 — Guard It at the Capability-Pair Level
Opus 4.7 fast mode is being retired on 2026-07-24. The model ID stays valid while only the speed parameter starts failing, so model-ID audits miss it. Here's a capability-pair preflight with automatic migration.
How Many Concurrent Claude API Requests Can You Actually Hold? Sizing Production Infrastructure with Little's Law and Measured Memory
Concurrency, queue depth, and memory are numbers you can derive, not guess. A working method for sizing Claude API production deployments with Little's Law, a memory probe, and a 30-minute load check — learned the hard way from an OOM crash.