[BUG] Every command retries up to 10 times ("Retrying in Xs · attempt X/10") and never gets a response. Happens on every single input.

Found any solution?

Same here, In my case i had to remove "CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS": "1" from my ~/.claude/settings.json to make it work again

Still same issue on every computer...

Same here. suddenly "Retrying in 0s · attempt 3/10"

For people who have issue as of now when I am commenting. Claude server is having some problem. It will be resolved soon. You can check status at https://status.claude.com/

I've had this issue few days already.. few prompts working but slow and then giving this error..

having this problem currently, has anyone found a way to fix this?

Hey, so i failed to make it work for claude code, but in my case i was not using paid calude code. i was using free model from open router & the issue was that when i configured the model, i forgot to update it as global.

If your case is same as mine, the fix is quite simple:

1. Open where claude code is installed in pc.
2. Locate setting file
3. Replace the file within claude directory with the file on which you have setup openrouter model.
4. Open a random project from rndom directory & type claude in terminal. You will see the imported model name there written.

Happy vibe coding.

Still having this issue :(

share your settings file draft,
here is what i used:

{
"env": {
"ANTHROPIC_BASE_URL": "https://openrouter.ai/api",
"ANTHROPIC_AUTH_TOKEN": "",
"ANTHROPIC_API_KEY": "",
"ANTHROPIC_MODEL": "minimax/minimax-m2.5:free"
}
}

I had same issue, narrowed it down to the MCP that was causing it. Check if you have MCP running and if so, turn it off and try.

Because direct auth checks can still look fine while Claude Code keeps retrying, I would separate this into three layers instead of treating it as one generic network problem:

1. Upstream reachability

• direct curl to the same base URL
• same shell, same env vars, same proxy settings

1. Local config merge

• project settings vs global ~/.claude/settings.json
• stale ANTHROPIC_BASE_URL / ANTHROPIC_MODEL / auth env vars
• any experimental flags that change request behavior

1. Runtime extensions

• MCP servers
• background hooks
• anything that adds latency or breaks the request loop after startup

The fact that some people fix it by removing an experimental flag, others by disabling MCP, and others by correcting a global base URL tells me there are multiple retry causes being grouped under the same symptom.

If I had to isolate this fast, I would test in this order:

1. Start from a clean shell with only the minimum auth/base-url env vars.
2. Temporarily move ~/.claude/settings.json aside and run with no MCP.
3. If using an OpenAI-compatible or OpenRouter-style base URL, confirm the config is global where Claude Code actually reads it.
4. Compare one failing launch with a run that has MCP and experimental flags disabled.

That usually shows whether the retries come from transport, settings merge, or an extension/runtime layer.

If anyone needs a quick paid BYO-key config/debug pass for Claude Code retry loops on Windows or custom base URLs, I can help isolate the exact config conflict and turn it into a stable working setup.

Closing for now — inactive for too long. Please open a new issue if this is still relevant.