Blog & Guides

Practical LLM API guides, model reviews, and integration tutorials.

Model reviewsNov 4, 2026

Tencent Hunyuan Hy4: WorkBuddy vs CodeBuddy Use Cases

Hy4's formal release adds visual understanding and computer operation. Learn when to use WorkBuddy for office tasks versus CodeBuddy for coding, plus API access via JZS Token.

Model reviewsOct 27, 2026

OpenRouter alternative: a developer's comparison

OpenRouter vs JZS Token: per-token pricing, routing, privacy, model count, free credit, and migration effort, compared for developers.

Model reviewsOct 27, 2026

OpenRouter alternatives in 2026: four options, compared

Four OpenRouter alternatives for 2026 — JZS Token, LiteLLM, direct provider APIs, and staying put — with trade-offs and pricing.

Model reviewsOct 27, 2026

Cheaper than OpenRouter? Here is the actual math

JZS Token vs OpenRouter prices on five current models, with 100M-token monthly savings math and the cases where OpenRouter is cheaper.

API and integration guidesOct 27, 2026

How to migrate from OpenRouter

A step-by-step OpenRouter migration guide for OpenAI-compatible apps: base URL, model-ID mapping, a test request, common errors, and a safe dual-run cutover.

Model reviewsOct 27, 2026

Cheapest LLM API in 2026: what the price list does not tell you

The cheapest LLM API options in 2026, ranked by blended price per 1M tokens, with the trade-offs for Claude, GPT, DeepSeek, and Qwen.

Model reviewsOct 27, 2026

LLM API pricing comparison 2026: 12 models, two gateways

2026 LLM API cost comparison across GPT, Claude, DeepSeek, Qwen, Gemini, and GLM, with input, output, and 3:1 blended prices.

API and integration guidesOct 27, 2026

How to use Claude Code with a custom API (your own key)

Configure Claude Code with a custom Anthropic-compatible API: env vars, the base-URL gotcha, model selection, and the errors people hit first.

API and integration guidesOct 27, 2026

How to use Cursor with a custom API (your own key)

Set up Cursor with a custom OpenAI-compatible API: the override base URL, model IDs, a test request, real per-token pricing, and Cursor's limits.