Run your coding agent against any model. One command.
Point Claude Code, Codex, or Grok Build at the AnyRouter gateway — one key, automatic failover, no shell edits. First run sets you up; every run after drops straight into your tool.
Copy and paste into terminal to try
v0.3.0 · Unified gateway for your coding tools · Claude Code
- Tool
- Claude Code
- Model
- auto
- Preset
- @preset/coding-stack
- API key
- sk-ar••••
› Select model / preset ‹ auto ›
Effort level ‹ high ›
Load hub ‹ none ›
▶ Open Claude Code
01Launchers
Pick your agent
One CLI wraps the agents you already use. The same stored key powers every launcher — copy a command and run it, no install step.
Claude Code
Anthropic-compatible
Launch Claude Code through AnyRouter. Failover, logging, and cost tracking come for free.
Codex
OpenAI-compatible
Drive OpenAI Codex against any model in the catalog. Set the effort level right from the menu.
Grok Build
xAI Grok CLI
Run xAI's Grok Build with the same one-key setup and automatic fallback on rate limits.
02Skills & Knowledge Hub
Load a hub straight into a session
A hub is your agent's context — skills, prompts, knowledge, and plugins — managed in the dashboard and pulled to any machine. Point a launcher at one and it syncs, shows you exactly what landed, then starts your agent with that context already in place.
Pick a hub — or let it list them
Pass --hub my-hub to load one directly, or run a launcher on its own and the CLI lists every hub you can see so you can pick one interactively.
Every item carries a scan verdict
Each skill, prompt, and doc is scanned for prompt injection before it can sync. The CLI prints exactly what landed — flagged items never reach your machine.
Loaded with --plugin-dir
The hub syncs to ~/.anyrouter/hubs/<slug> as a self-contained Claude Code plugin, then launches with --plugin-dir pointed at it. Your ~/.claude/skills is never touched.
Skills load, namespaced
Claude auto-discovers the hub's skills as anyrouter-my-hub:<skill> — your agent starts with that context already in place.
One hub, shared with your whole team
Build a hub once in the dashboard — skills, prompts, knowledge, and plugins versioned together. Teammates pull the same context, and you can publish it read-only for everyone.
my-hub
28 items · synced & scanned · v14
You
Pull it onto any machine — --hub my-hub and your agent starts with full context.
Your team
6 teammates run the exact same skills & knowledge — no copy-paste, no drift.
Everyone
Publish read-only at anyrouter.dev/h/my-hub — anyone can load it.
03Guided menu
Interactive when you want it
Run a launcher with no flags and you get a guided menu — preview config, change model or effort, load a hub, then launch. Pass flags and it skips straight through.
Pick model or preset
Arrow through 122+ models or a saved @preset — no slug to memorize.
Tune effort inline
Set reasoning effort from low to xhigh before the agent starts.
Attach a hub
Load your skills & knowledge for this session, right from the menu.
All keyboard, no mouse
↑↓ to navigate, ←→ to change, Enter to launch. Scriptable too.
v0.3.0 · Unified gateway for your coding tools · Claude Code
Configuration
- Tool
- Claude Code
- Model
- auto
- Preset
- @preset/coding-stack
- API key
- sk-ar****
04Under the hood
How it works
Three steps, once. After that, every launch goes straight into your tool.
Sign in once
Paste an sk-ar- key or log in through the browser. Your profile is saved locally and reused by every launcher — no re-auth per tool.
Pick any model
Choose with --model or a preset from 122+ models. AnyRouter handles auto-failover on rate limits and provider errors.
Launch the real agent
The CLI injects base URL, auth, and model into the child process only, then spawns the real claude / codex / grok.
Your config follows you
Keys, presets, and hubs live in AnyRouter and load on any machine. Sign in and your whole setup is just there.
Clean exit, every time
Credentials live only in the agent's process. When it exits, they're gone — no env vars or files left behind.
05Models
Swap models in one flag
No provider accounts to rotate, no settings.json edits. Change --model and AnyRouter handles routing, failover, and billing.
# Swap models without touching confignpx @anyr/cli claude --model z-ai/glm-5.2# Let a preset pick the model, set reasoning depthnpx @anyr/cli claude --model auto --effort high# Same key works for Codexnpx @anyr/cli codex --model anthropic/claude-sonnet-4.6 --effort xhigh
Per-project defaults live in anyrouter.yaml — check it in, and the whole team gets the same routing, effort, and hub. See the configuration reference.
# defaults for every launcherdefault: @preset/coding-stackeffort: highhub: my-hub# per-launcher overridesclaude:model: anthropic/claude-sonnet-4.6codex:model: z-ai/glm-5.2effort: xhigh# auto-failover on these statusesfallback: [ 429, 401, 5xx ]
06Documentation
Everything else is in the docs
Guides for each launcher, the hub, configuration, and security — all in one place.
CLI overview
Interactive menu, first-run setup, shared config, management commands.
Claude Code
One-command setup, model flags, and manual env-var fallback.
Skill & Knowledge Hubs
Sync shared skills, prompts, and plugins to every machine with one command.
Configuration
anyrouter.yaml, profiles, JSON schema, and all CLI flags.
Login & security
Browser PKCE flow, device code flow, and key storage.
Model catalog
Browse every model id you can pass to --model.
Try it in your terminal
Free to start. Paste a key or log in on first run — then every coding session routes through AnyRouter.
Copy and paste into terminal to try