Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchYou can use Codex with your personal ChatGPT sign-in or configure it to send supported model requests through Amazon Bedrock—but Bedrock is not a second account in ChatGPT’s account switcher. Switching to Bedrock means changing Codex’s model-provider configuration and using AWS authentication, permissions, Regions, and billing. To return to your personal ChatGPT-backed setup, restore its provider configuration; the ChatGPT account switcher itself applies to ChatGPT web, not Codex desktop.
What “switching” means in Codex
There are two separate mechanisms. ChatGPT’s account switcher lets you keep two ChatGPT accounts signed in and switch between them; the accounts remain independent. OpenAI says this switcher is available on ChatGPT web and is not supported in Codex desktop or native ChatGPT mobile apps. It does not add Amazon Bedrock to the ChatGPT account menu. OpenAI’s account-switching guide explains the ChatGPT feature.
As an Amazon Associate I earn from qualifying purchases.
For Codex, Bedrock is a model-provider choice. You configure Codex to send supported requests to Amazon Bedrock, where AWS handles credentials, permissions, model access, quotas, Regions, and billing. The Codex client can remain on your computer in either setup, but the provider changes how requests are authenticated and handled, and which features are available. OpenAI’s Bedrock overview describes the boundary.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →What changes between the two setups
| Area | Personal ChatGPT-backed Codex | Codex through Amazon Bedrock |
|---|---|---|
| What you change | ChatGPT sign-in or account context | Codex’s local model-provider configuration |
| Authentication | Your existing ChatGPT authentication | A Bedrock API key or AWS SDK credentials |
| Billing and administration | Your ChatGPT account and plan context | AWS account billing, IAM permissions, model access, quotas, and Region availability |
| Feature support | Depends on your ChatGPT and Codex configuration | Some OpenAI-hosted cloud features are unavailable; API capabilities vary by model, endpoint, Region, and account configuration |
| Support ownership | OpenAI for Codex client behavior | OpenAI for Codex client setup; AWS or your AWS administrator for credentials, permissions, access, quotas, billing, Regions, and Bedrock service behavior |
Codex can keep local files and conversation history in either setup, but continuity should not be assumed for every system or version. In an AWS-published tutorial, Matheus Guimaraes reported that his conversations and generated files remained visible when he changed provider configuration on macOS; that is an individual test, not a universal guarantee. Read the tutorial and its account of that test.
#1 Best Overall
Configure Codex to use Amazon Bedrock
1. Check model and Region access
Choose a model ID supported in the AWS Region you plan to use, and confirm that your AWS account can access it. Model availability and access requirements differ. AWS says many foundation models are enabled by default when the necessary AWS Marketplace permissions are in place, while some require additional account-level access or prerequisites. AWS’s model-access guide explains the process.
2. Choose how Codex will authenticate
Codex supports a Bedrock API key through AWS_BEARER_TOKEN_BEDROCK, or the AWS SDK credential chain, which can use sources such as shared AWS configuration, environment variables, AWS SSO, and named profiles. If the Bedrock API-key variable is set, Codex checks it before falling back to SDK credentials. Do not use OPENAI_API_KEY for the Bedrock provider.
For desktop or IDE use, do not assume the app inherits environment variables from your shell. OpenAI documents putting environment variables in ~/.codex/.env for this situation. Keep credentials out of shared or public files.
3. Set the provider, model, and Region
Edit ~/.codex/config.toml and set the Bedrock provider, a supported model ID, and a Region. OpenAI’s example uses the following fields; treat the model ID as an example rather than a universal choice, and use one supported by your account and selected Region:
model = "openai.gpt-5.6-sol"
model_provider = "amazon-bedrock"
[model_providers.amazon-bedrock]
wire_api = "responses"
region = "us-east-1"The API-key path needs a Bedrock Region in Codex configuration. With SDK credentials, set an explicit Region or make sure the AWS SDK can resolve one from your configuration, environment, or profile. See OpenAI’s configuration guide for the documented setup and options.
4. Restart and verify
After changing ~/.codex/config.toml or ~/.codex/.env, restart the Codex desktop app or VS Code extension. In the CLI, use /status to verify the active configuration; in desktop or IDE, start a new session after restarting.
Switch back to your personal ChatGPT-backed setup
Restore the provider configuration that Codex used for your personal ChatGPT-backed setup, then relaunch Codex. The ChatGPT web account switcher does not perform this provider change. If you manually edited ~/.codex/config.toml, back it up first and remove or change only the Bedrock settings you added; avoid replacing unrelated configuration.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Clear out junk files and repair common Windows errors3Fix the driver behind crashes, sound loss and screen glitchesThe AWS-published tutorial linked above describes one local approach: remove its specifically marked Bedrock configuration block, then relaunch the desktop app. Its author reported that his prior local conversations and generated files remained available on macOS, but that result should not be treated as a guarantee for other platforms or versions. The tutorial also provides optional codex-bedrock and codex-personal shell commands. Those commands come from an independent third party, not from an OpenAI-supported account-switch feature; inspect any script and make a backup before running it.
If a Bedrock request fails
- Check the provider and model. Confirm
model_providerisamazon-bedrock, the model ID is exact and supported, and the request is configured for Bedrock rather than an OpenAI-hosted API. - Check the Region. Make sure Codex and the AWS SDK resolve the intended Region, and that the selected model and endpoint are available there.
- Check which credentials Codex sees. A set
AWS_BEARER_TOKEN_BEDROCKtakes precedence over SDK credentials. If a desktop or IDE app cannot see shell variables, check~/.codex/.env. - Check AWS access. Verify the active AWS identity, IAM permissions, model access, any model-specific prerequisites, and quotas with your AWS administrator or AWS.
- Check API capability. A capability described for an OpenAI-hosted API may not be available through Bedrock. Support depends on the model and Bedrock endpoint; consult OpenAI’s Bedrock API guide.
- Restart and start a new session. Restart the app or extension after configuration edits, then test in a new session.
OpenAI Support is the contact for Codex client setup and local behavior. AWS or your AWS administrator is the appropriate contact for AWS credentials, IAM, model access, quotas, billing, regional availability, and Bedrock-side request failures.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
