One API for all your AI agents
Works with OpenAI- and Anthropic-compatible agents like Claude Code, Codex and WorkBuddy. One account, every charge itemized.
Already have an account? Sign in
- Channels
- 6+
- Top-tier models
- 25+
- Provider
- 50+
- Uptime SLA
- 99.5%
- agent
- claude-code
- endpoint
- codelux.io/v1/messages
- route
- anthropic / claude-sonnet-5
- status
- available
- claude-sonnet-5
- claude-opus-5-5
- gpt-6-sol
- DeepSeek-V41-Flash
- kimi-k3
- glm-5.3
Connect your agent in three steps
Every agent shares one Codelux account and bill.
Pick your agent
Open its channel for setup notes.
Create an account
Register, then pick a plan or top up as you go.
Configure in Portal
Create an API key and add it to your agent.
- Claude CodeAnthropic's official agent: fix bugs, write tests, run migrationsSetup
- CodexOpenAI's official agent: run multiple tasks in parallelSetup
- OpenCodeOpen-source AI coding agent: use models from any providerSetup
- Kimi CodeMoonshot's official agent: Kimi K3 with up to 1M contextSetup
- DeepSeek HarnessDeepSeek's open-source agent framework: everything is a pluginSetup
- Tencent WorkBuddyBuilt by Tencent: a new way to work with AI agentsSetup
Public model pricing
Prices are per 1M tokens. Sign in for member prices.
Loading public pricing…
Subscription plans
Loading plans…
Prefer not to subscribe? Top up any amount and pay only for what you use. Top-up options and first top-up offers are shown in Portal. Sign in to top up
What users say
I used to keep separate accounts for Claude Code and Codex. Now one key covers both, and there's a single bill at the end of the month.
FAQ
Which agents are supported?
Any agent that lets you set a custom API endpoint and speaks an OpenAI- or Anthropic-compatible API. These have step-by-step setup channels: Claude Code, Codex, OpenCode, Kimi Code, DeepSeek Harness, Tencent WorkBuddy.
How do I connect an agent to Codelux?
Register, create an API key in Portal, then put the Codelux endpoint and key into your agent's custom API settings. Claude Code can be imported from Portal in one step with CC-Switch.
How are model prices calculated?
Input, output, cache write and cache read are each priced per 1M tokens, in the currency shown in the table.
Where can I see member pricing?
Sign in and open the model hub in Portal.
How do plans differ from pay-as-you-go?
A plan is paid once; its credit lands in its own wallet, is used before your pay-as-you-go balance and expires at the end of its term. Pay-as-you-go needs no subscription and charges usage to your balance.
Does the first top-up offer stack with plans?
No. Plans are priced on their own, and buying one keeps your first top-up offer available for pay-as-you-go.
Where can I see usage and charges?
Sign in to Portal and open Logs to see the model, token usage and charge for each request.
How do I get help?
Sign in to Portal and open a ticket.
Start with one account
After registering, create API keys and choose a plan or top up in Portal.
Already have an account? Sign in