Every leading model inside
Codex
and
Claude Code. AutoRouter sends each task to the
model that does it best, so you spend up to 50% less and stop
thinking about models.
Free with your ChatGPT or Claude plan · macOS
Codex
Claude Code
OpenAI
Anthropic
Google
xAI
DeepSeek
Moonshot
Z.ai
MiniMax
No new app to learn and nothing to configure. AutoRouter shows up in the model picker you already use.
A small Mac app that connects to the plans you already have.
Pick AutoRouter from the model menu in Codex, or with
/model in Claude Code.
Each turn goes to the model that fits it, and you stop wasting credits on the wrong models.
New models show up in your picker right away, and AutoRouter sends them the work they do best. Or override with a specific model any time.
AutoRouter understands how hard each task is and sends it to the model that fits.
Codex
Refactor the billing module to use the new ledger API
AutoRouterClaude Opus 5.5Claude plan
GPT-6 AstraChatGPT plan
MiniMax M3.1 FlashPro
Examples. Real routes follow your plans and rules.
Every model has to clear a quality bar for the task before price counts at all. Hard work goes to the strong models on your plans.
Ratings retune the router constantly, along with tests of every new model as it becomes available. Benchmarks and automatic evaluations on real coding tasks are coming soon.
Your rating
Plan the ledger migration
Next turn
Plan the payouts migration
Retuned from ratingsSend UI prompts to frontier models and git commands to a free one. Take your intelligence toolkit with you.
Claude Opus
5.5
for UI work.14 turns this week
GPT-6.1
Sol
for code reviews.9 turns this week
MiniMax M3.1
Flash
for git commands.31 turns this week
Write a skill once and everyone on your team can use it in every harness. No more naming conflicts or out-of-date copies.
release-notesDrafts release notes from merged PRs
Only you
code-reviewReviews a PR against your standards
Everyone
db-migrationPlans a migration and its rollback
Platform team
incident-timelineBuilds a timeline from your alerts
On-call
design-tokensTurns your design tokens into UI
Design team
Your plans go to the work that needs them, so they last longer. No subscription? Credits with Pro get you every model.
An example week of 30% hard, 45% medium and 25% quick turns,
at
Claude Opus, Sonnet and Haiku 5.5 list prices with cached input at a
tenth of input. Pro sends 85% of quick and 75% of medium turns
to open models, which its included usage pays for.
Every developer gets the models they need. You get rules, budgets and a clear view of what each change costs.
Routing across your
ChatGPT
and
Claude
plans is free. Pro adds every other model, with $20 of included
usage each month.
Your subscriptions only.
$0/mo
Free to route across your
subscriptions
Codex
and
Claude
Code
Every model and provider
$20/mo
$20 of included usage each month*
Plus everything in Free
Start ProRolled out to your whole team.
Custom
Priced by seats for your team
Plus everything in Pro
Get in touch*Included usage resets each month and doesn’t roll over. New accounts get Pro free for their first month, or until $20 of usage is spent, whichever comes first. Usage is billed at the provider’s price plus 10%; credits you top up never expire.
Quality, routing, billing and privacy. Anything else, email support@autorouter.com.
A Mac app that adds 20+ models to Codex and Claude Code, plus one more choice: AutoRouter, which picks the right model for each turn. It uses the ChatGPT and Claude plans you already pay for first.
We built it so they don’t. Every model has to clear a quality bar for the task before price counts at all, unless one of your own rules names it, and hard work goes to the strongest model your plans cover, even when a cheaper one comes close. If a pick ever falls short, rate it, or choose a model yourself.
You can see which model handled every turn and what it cost, change it with a rule or your app’s model picker, and use models from every provider, including your other plan.
Your rules go first: if one applies, it decides the turn. Otherwise AutoRouter works out how hard the turn is and only considers models that clear the bar for it. Hard work goes to the most capable model your plans cover, and to credits only when none is strong enough. Everything else goes to the one that costs you least, which is often your plan or a low-cost open model.
Yes. AutoRouter’s models appear in your app’s model picker beside the ones you already have. Choose any of them and your turns go straight to it; AutoRouter routes only while it’s the model selected.
Yes. Our team reviews real routing decisions and every rating you send, and retunes AutoRouter with them. Benchmarks for every model on real coding tasks, and routing that learns on its own, are coming next.
Yes, paid models are billed at provider price plus 10%. But price only decides between models that already clear the bar, and your plans come first. We earn nothing on them.
It runs on macOS 14 or later, with Codex and Claude Code; more apps are coming. It sometimes misjudges how hard a task is, so rate the pick when it does. We don’t publish benchmarks yet, and a shared skills library for every tool and your whole team is coming soon.
Codex, in the app and the CLI, and the Claude Code CLI. More apps are on the way. AutoRouter runs on macOS 14 or later.
It adds AutoRouter to each app’s own settings, Codex’s
config.toml and Claude Code’s
settings.json, after saving a copy of each
file. Your other settings stay as they are. Disable an app
in AutoRouter’s settings to switch it back.
Yes, first. In Codex, OpenAI models run on your ChatGPT plan; in Claude Code, Claude models run on your Claude plan. Connect the other subscription in settings to use your Claude plan inside Codex, or your ChatGPT plan inside Claude Code.
No. Pro gives you every model, with $20 of included usage each month, and you can top up anytime.
When your ChatGPT or Claude plan hits its limit, Pro keeps those turns going from your monthly usage until the plan resets. If that runs out too, AutoRouter asks you to top up.
Routing across your ChatGPT and Claude subscriptions is free. Every other model comes with Pro: $20 a month with $20 of included usage each month. Paid models are billed at the provider’s price plus 10%, from that usage first and then from credits you top up. New accounts get Pro free for their first month, or until $20 of usage is spent, whichever comes first.
Requests pass through AutoRouter’s servers on their way to each provider, and we don’t store transcripts. We keep the message that opens each turn, up to 4,000 characters, to review routing, along with records of the model chosen, tokens and cost. The Privacy Policy has the details.
Everything in Pro, plus rollout through your MDM; model and budget rules for each team; spend by team, project or ticket; zero data retention; and SSO, SCIM and an audit log. Spend by pull request is coming soon. Get in touch to roll it out.
Download AutoRouter, choose it as the model in
Codex
or
Claude Code, and keep working the way you do now.
Free with your ChatGPT or Claude plan · macOS
Three steps to get started.