Blog
Notes from the routing layer.
Practical guides to model choice, routing economics, and working across AI providers in Codex Desktop.
How to use Codex in 4 steps with 1routers
Set up Codex through 1routers with real screenshots: download the right app, connect your account, and send a safe first task without editing config files.
Read article
How to use DeepSeek, Qwen, and Kimi in Codex with 1routers: a 4-step beginner guide
Use DeepSeek, Qwen, and Kimi in Codex through 1routers without configuring a separate provider API key for every model family.
Read articleHow beginners can use MCP in Codex: build a persistent cat memory in 5 steps
Install a local Memory MCP, store a fictional cat profile, retrieve it in a new Codex task, and turn those facts into a small birthday plan—no coding required.
Read articleHow beginners can use a Codex Skill: make a Chinese PowerPoint in 5 steps
Install the Presentations plugin, copy one prompt, and create an editable six-slide cat café PowerPoint with Codex—no coding required.
Read articleHow beginners can use Codex Tools: organize 9 messy weekend photos automatically
Download nine practice photos, copy one safe prompt, and let Codex group, rename, and index them without touching the originals or writing code.
Read articleCodex costs too much? 7 practical ways beginners can spend less
Reduce Codex model costs with seven beginner-friendly changes to model choice, routing tier, context, retries, and usage review.
Read articleBuilt around real work
Start from tasks, routes, and observed usage—not keyword volume.
Answer first
State the useful answer before expanding into limits and steps.
Maintained over time
Update details as models, rates, and availability change.