prodex
prodex is a wrapper for Codex and Claude Code for working with multiple profiles.
The main feature is auto rotate. If one OpenAI/Codex profile runs out of quota, prodex can route new work to another profile that is still available. You do not need to switch accounts manually.
Why use it
Use prodex if you want to:
- automatically use another available profile when quota runs out
- work with multiple accounts
- keep each profile isolated
- keep sessions tied to the original profile
If you only use one account and do not need profile isolation or quota-aware routing, you probably do not need it.
Requirements
You need at least one logged-in Prodex profile.
Depending on your setup, you may also need:
- Codex CLI for
prodexandprodex caveman - Claude Code for
prodex claude claude-memformemvariants
Install
npm
Cargo
If you install with Cargo, make sure the codex binary in your PATH is already installed and up to date.
Quick start
If your current Codex home is already logged in:
Or set it up from scratch:
You can also import a logged-in Copilot CLI account:
Check your profiles and quota:
Run through prodex:
Common commands
Run Codex
|
Run Caveman mode
prodex caveman runs Codex with a temporary overlay CODEX_HOME, so the base profile home stays unchanged after the session ends.
If you use the mem variant, Prodex points an existing Claude-Mem Codex setup to the active Prodex session path instead of the default ~/.codex/sessions.
Run Super mode
prodex super is a shortcut for prodex caveman mem --full-access.
Use this when you want Caveman mode, Claude-Mem transcript watching, and launch-time full access together. Full access maps to Codex's sandbox-bypass launch flag, so use it only when you intentionally want Codex to run without the normal approval and sandbox protections.
Use prodex super --url http://127.0.0.1:8131 when you want the same Super mode front end to talk directly to a local OpenAI-compatible server such as llama-server. Prodex injects a temporary prodex-local Codex provider, appends /v1 when the URL has no path, disables non-function native tools that local servers commonly reject, advertises a conservative 16k local context window, and skips quota/proxy routing for that launch. The default local model id is unsloth/qwen3.5-35b-a3b; override it with --model, for example prodex super --url http://127.0.0.1:8131 --model local/qwen. Use --context-window and --auto-compact-token-limit if your local server is configured larger. See LOCAL.md for self-hosted model setup and testing.
Add --dry-run to run, Caveman, or Super launches to print the resolved provider, model, CODEX_HOME, proxy args, and launch env with secret-looking values redacted. Dry-run output is prelaunch only and does not start Codex or the TUI.
Run Claude Code
prodex claude uses the normal Claude Code flow.
prodex claude caveman enables Caveman for that session while keeping state under the Prodex-managed CLAUDE_CONFIG_DIR, not the global ~/.claude.
prodex claude caveman mem combines Caveman and Claude-Mem.
Profile commands
Bedrock and custom providers
Auto rotate and quota checks apply to supported OpenAI/Codex profiles.
If a profile's config.toml sets model_provider to a non-OpenAI backend such as amazon-bedrock, prodex run and prodex caveman launch Codex directly without quota preflight or the local auto-rotate proxy. Bedrock quota, credentials, regions, and provider errors are handled by Codex and the upstream provider, not by Prodex.
prodex quota is not available for those profiles.
prodex claude is only supported with the default OpenAI/Codex provider.
When you import a Copilot profile, Prodex does not move the Copilot token into Prodex-managed storage. It only records the provider identity and API endpoint in its own metadata.
Utility commands
More
See QUICKSTART.md for a longer walkthrough, and LOCAL.md for self-hosted local model setup.
Contributor testing guidance lives in docs/testing.md, including the fast/serial split and runtime parallel-safety assumptions.