xtekky/gpt4free · error · ResponseError

result["error"].get("message", result["error"])

Error message

result["error"].get("message", result["error"])

What it means

Raised as ResponseError while consuming Puter's text/event-stream response: an SSE event arrived whose JSON payload contains an 'error' key. The provider surfaces the upstream error message (or the raw error object) and terminates the stream.

Solutions

1. Catch ResponseError and inspect the message; fix the model name or request payload it complains about
2. Check Puter account quota/usage if the message indicates rate or credit limits
3. Retry with backoff for transient upstream errors; fall back to another provider if persistent

Example fix

# before
for chunk in client.chat.completions.create(..., stream=True):
    print(chunk.choices[0].delta.content)

# after
try:
    for chunk in client.chat.completions.create(..., stream=True):
        print(chunk.choices[0].delta.content)
except g4f.errors.ResponseError as e:
    logging.error('Puter stream error: %s', e)
    raise

from g4f.errors import ResponseError
try:
    stream = ...create(..., stream=True)
    for chunk in stream: process(chunk)
except ResponseError as e:
    if 'rate' in str(e).lower():
        backoff_and_retry()
    else:
        raise

When it happens

Trigger: Streaming a completion from Puter when the upstream service returns an error event mid-stream: invalid model, quota exceeded, rate limit, or malformed request accepted late.

Common situations: Using a model name Puter no longer supports, hitting account rate limits partway through a long streaming session, transient upstream 5xx surfaced as SSE error events.

Related errors

• line[6:]
• result
• API key is required for Puter.js API
• Expected SSE response but got content-type
• Invalid JSON data

AI-assisted analysis of xtekky/gpt4free@973504e177 (2026-08-14). Data as JSON: /api/errors/7f71be778d2a3baa. Report an issue: GitHub.

Appendix: source

Thrown at g4f/Provider/needs_auth/Puter.py:486

                        param: kwargs.get(param)
                        for param in extra_parameters
                        if param in kwargs
                    },
                },
            }
            async with session.post(
                cls.api_endpoint, headers=headers, json=json_data, proxy=proxy
            ) as response:
                await raise_for_status(response)
                mime_type = response.headers.get("content-type", "")
                if mime_type.startswith("text/plain"):
                    yield await response.text()
                    return
                elif mime_type.startswith("text/event-stream"):
                    reasoning = False
                    async for result in sse_stream(response.content):
                        if "error" in result:
                            raise ResponseError(
                                result["error"].get("message", result["error"])
                            )
                        choices = result.get("choices", [{}])
                        choice = choices.pop() if choices else {}
                        content = choice.get("delta", {}).get("content")
                        reasoning_content = choice.get("delta", {}).get(
                            "reasoning_content"
                        )
                        if reasoning_content:
                            reasoning = True
                            yield Reasoning(reasoning_content)
                        elif content:
                            if reasoning:
                                yield Reasoning(status="")
                                reasoning = False
                            yield content
                        if result.get("usage") is not None:
                            yield Usage(**result["usage"])

View on GitHub (pinned to 973504e177)