博客

Guide

DeepSeek in Claude Code: Official Model Names and Relay Setup Checks

Use DeepSeek’s official request name, distinguish legacy aliases from relay catalog IDs, and check Claude Code protocol, key type and billing before connecting.

2026-06-30 4

LLM API Editorial Team

Guide

Anthropic API Rate Limits & Usage Tiers Explained (2026)

How Claude API rate limits work in 2026: RPM, ITPM, and OTPM; the four spend-based usage tiers and how you advance; why cached reads don't count; and how to handle 429s.

2026-06-29 8

Yazhou Hu

Tutorial

ANTHROPIC_BASE_URL: Claude Code Setup and /v1 Fixes

Set ANTHROPIC_BASE_URL correctly for Claude Code on Windows, macOS and Linux. Compare base URLs with request endpoints and diagnose 404, authentication and quota errors.

2026-06-28 5

LLM API Editorial Team

Comparison

Cheap Claude API Access in 2026: The Legit Ways to Pay Less

The real ways to get cheaper Claude API access in 2026 — right-sizing the model, prompt caching, batch, subscription vs pay-as-you-go, and relays — with an honest look at the trade-offs of each.

2026-06-27 8

Yazhou Hu

Guide

Claude Code Token Cost: Why It Climbs and How to Cut It

Claude Code token cost grows super-linearly because context is re-sent every turn. Here's why long sessions get expensive and the concrete levers to cut spend.

2026-06-23 7

Yazhou Hu

Guide

What Is a Claude-Compatible API Relay? A Plain Explainer

A claude compatible api relay speaks the same endpoints Claude Code already uses. Here's how relays work, why your CLI can't tell, and whether to trust one.

2026-06-16 7

Yazhou Hu

Comparison

Claude Sonnet 4.6 vs Opus 4.8 in Claude Code: Which to Run

Claude Sonnet 4.6 vs Opus 4.8 for Claude Code: Sonnet is cheaper and faster for daily coding; Opus 4.8 wins hard refactors and architecture. Task-by-task picks, token cost, and the in-session /model switch.

2026-06-09 7

Yazhou Hu

Comparison

Claude Code vs Codex CLI in 2026: An Honest Head-to-Head

Claude Code vs Codex CLI compared on workflow, protocol, and config — run both agentic coding CLIs off one sk-relay key because we expose both protocols.

2026-06-06 8

Yazhou Hu

Guide

Claude Opus 4.8 in Claude Code: Setup, the 1M Context Window, and What It Costs

Run Claude Opus 4.8 in Claude Code: exact model IDs, the ANTHROPIC_BASE_URL setup, whether the 1M-token context window is still available, and how to keep the bill down.

2026-06-06 7

Yazhou Hu

Troubleshooting

Fix a Claude Code Connection Error or Timeout: A Diagnostic Ladder

A Claude Code connection error or timeout usually has a boring cause. Here's the diagnostic ladder to tell a real network/config fault from the model thinking.

2026-06-06 7

Yazhou Hu

Comparison

Claude Code Pricing in 2026: Pro vs Max vs Pay-As-You-Go API

A plain-English breakdown of what Claude Code actually costs in 2026 — the official Pro and Max subscription tiers, how token-metered API billing compares, and when a relay subscription at a fraction of the price makes sense.

2026-06-05 8

Yazhou Hu

Tutorial

Using Cline and Roo Code with a Claude-Compatible API

Step-by-step setup for pointing Cline and Roo Code — the VS Code AI coding extensions — at an Anthropic-compatible relay. Provider settings, base URL, model IDs, and the gotchas with OpenAI-compatible vs native Anthropic mode.

2026-06-04 7

Yazhou Hu

Announcement

Codex CLI, OpenAI SDK & LiteLLM with Claude: OpenAI-Compatible Setup

Point Codex CLI, the OpenAI SDK, or LiteLLM at Claude models with one OpenAI-compatible key. Exact /v1/responses and /v1/chat/completions setup, the wire_api gotcha, and a Codex config.toml walkthrough.

2026-05-15 6

Yazhou Hu

Engineering Notes

A 15-Second CPU Stall From One Unclosed XML Tag

Postmortem on a relay process that pegged a single core for fifteen seconds straight. Catastrophic regex backtracking on user-controlled input. Three-character fix, three-layer defense (bounded quantifier + per-request watchdog + nginx next_upstream).

2026-05-06 9

Yazhou Hu

Engineering Notes

Reasoning Blocks at the Relay: What Changes When the Upstream Emits Thinking

Open-weight reasoning models (DeepSeek V3.x, MiniMax M-series, Qwen3 reasoning) now ship interleaved thinking blocks. Notes on translating that to Claude Code's Anthropic-protocol contract — block ordering, signature handling, tool-use interleaving, and SSE shape.

2026-05-06 10

Yazhou Hu

Engineering Notes

Tool-Use IDs in the Wild: When the Same ID Appears Twice

Long Claude Code sessions sometimes produced duplicate `tool_use_id` values from the upstream model. Naive fixes break round-trip correlation; the working answer is occurrence-counter dedup with two-direction translation. With diagnostic dumps for the cases that still slip.

2026-05-06 8

Yazhou Hu

Engineering

Why Claude Code Sessions Break Past 400K Tokens — And the Self-Audit Pattern That Fixes It

Long Claude Code conversations degrade past ~400K tokens: fabricated paths, skipped TODOs, contradicted decisions. The naive fix (open a new window) fails because self-summaries lie. Here is the self-audit pattern that catches the lying — and the open-source skill that automates it.

2026-04-24 9

Yazhou Hu

Guide

Claude Code on a Budget: A Solo Developer Starter Guide

How to run Claude Code for free or under $5/month as a solo developer. Free tier tactics, when to upgrade, model selection, and the 2026 solo dev stack.

2026-04-23 9

Yazhou Hu

Guide

Running Claude Code from China: A Practical 2026 Guide

Real options for running Claude Code productively from mainland China: VPN tradeoffs, domestic-accessible relays, self-hosted proxies, and a 60-second setup with Alipay/WeChat billing.

2026-04-23 10

Yazhou Hu

Guide

Claude Code Tool Use: A Complete Developer Guide

Deep dive into Claude Code's six core tools (Read, Edit, Write, Bash, Grep, Glob), how the model decides which tool to call, parallel tool calls, and how to debug when tool use fails.

2026-04-23 11

Yazhou Hu