xtekky/gpt4free · critical · RetryProviderError

RetryProvider failed:\n{p}: {type(exception).__name__}: {exc

Error message

RetryProvider failed:\n{p}: {type(exception).__name__}: {exception}

What it means

RetryProviderError raised by raise_exceptions when two or more providers in a RetryProvider run each recorded an exception: the messages of every failure are combined ('provider: ExcType: message' per line) and chained from the first exception. It is the aggregated report of a fully failed multi-provider attempt; with exactly one failure, that original exception is re-raised instead.

Source

Thrown at g4f/providers/retry_provider.py:289

        else:
            async for chunk in super().create_async_generator(
                model, messages, **kwargs
            ):
                yield chunk


def raise_exceptions(exceptions: dict) -> None:
    """
    Raise a combined exception if any occurred during retries.

    Raises:
        RetryProviderError: If any provider encountered an exception.
        RetryNoProviderError: If no provider is found.
    """
    if exceptions:
        if len(exceptions) == 1:
            raise list(exceptions.values())[0]
        raise RetryProviderError(
            "RetryProvider failed:\n"
            + "\n".join(
                [
                    f"{p}: {type(exception).__name__}: {exception}"
                    for p, exception in exceptions.items()
                ]
            )
        ) from list(exceptions.values())[0]

    raise RetryNoProviderError("No content response from any provider. ")

View on GitHub (pinned to 973504e177)

Solutions

1. Read each line of the combined message — it enumerates every provider and its root cause; fix the most frequent cause (often blocks or missing auth).
2. Update g4f (pip install -U g4f) since mass provider failure usually means upstream sites changed.
3. Run from a residential IP / add valid credentials for at least one provider in the chain.
4. Narrow the chain to providers you can actually authenticate with, so one success short-circuits the rest.

Example fix

// before
result = await g4f.ChatCompletion.create_async(model, messages)  # default retry chain

// after
from g4f.errors import RetryProviderError
try:
    result = await g4f.ChatCompletion.create_async(model, messages)
except RetryProviderError as e:
    logging.error('all providers failed: %s', e)
    result = await my_authed_provider.create_async_generator(model, messages)

def is_retry_provider_error(e: BaseException) -> bool:
    from g4f.errors import RetryProviderError
    return isinstance(e, RetryProviderError) or str(e).startswith('RetryProvider failed:')

from g4f.errors import RetryProviderError
try:
    result = await g4f.ChatCompletion.create_async(model, messages)
except RetryProviderError as e:
    for line in str(e).splitlines():
        logging.warning('provider failure: %s', line)
    result = await authed_fallback(model, messages)

When it happens

Trigger: Iterating a RetryProvider (e.g. g4f's default fallback chain) where every provider throws — auth failures, blocks, rate limits — so the exceptions dict has 2+ entries when the generator completes without content.

Common situations: Free-provider outage waves where Cloudflare blocks and 429s hit every mirror simultaneously; expired cookies/credentials across the chain; running from datacenter IPs that most providers block; stale g4f version with broken provider implementations.

Related errors

• config.yaml: No provider succeeded for model {model!r}. Trie
• No providers available
• No content response from any provider.
• Polling images faild. Code: {response.status}
• GLM captcha solving failed after {SOLVE_RETRIES} attempts: {

AI-assisted analysis of xtekky/gpt4free@973504e177 (2026-08-14). Data as JSON: /api/errors/02d1c140b78e7a6c. Report an issue: GitHub.