xtekky/gpt4free · error · RuntimeError

PhindAi API returned success=False: {data}

Error message

PhindAi API returned success=False: {data}

What it means

Raised when PhindAi's ajax endpoint returns valid JSON but with success != true. The full data payload is embedded; typical reasons are an invalid nonce (WP ajax rejects the request), rate limiting, or the backend refusing the message content.

Source

Thrown at g4f/Provider/PhindAi.py:71

                    raise RuntimeError("Failed to extract nonce from PhindAi response")
                nonce = match.group(1)

            # 2. Fetch the response
            ajax_url = f"{cls.url}/wp-admin/admin-ajax.php"
            payload = {"action": "phind_ai_send", "nonce": nonce, "message": prompt}

            async with session.post(ajax_url, data=payload) as response:
                await raise_for_status(response)
                try:
                    data = await response.json()
                except json.JSONDecodeError:
                    text = await response.text()
                    raise RuntimeError(
                        f"Failed to decode JSON from PhindAi response: {text}"
                    )

                if not data.get("success"):
                    raise RuntimeError(f"PhindAi API returned success=False: {data}")

                response_text = data.get("data", {}).get("response", "")
                return response_text

View on GitHub (pinned to 973504e177)

Solutions

1. Read the embedded data field — it names the exact reason (nonce/rate limit/content).
2. Retry with a fresh session so a new nonce is fetched per request.
3. Slow down request cadence or rotate proxy.
4. Shorten/sanitize the prompt if the message content is being refused.

Example fix

# before - reusing one session/nonce for many prompts
async with StreamSession(...) as s:
    for p in prompts:
        ...

# after - new impersonated session per request
for p in prompts:
    async with StreamSession(headers=headers, impersonate='chrome') as s:
        await PhindAi.create_async_generator(model, [{'role':'user','content':p}])
    await asyncio.sleep(1)

try:
    result = await PhindAi.create_async_generator(model, messages)
except RuntimeError as e:
    if 'success=False' in str(e) and 'nonce' in str(e).lower():
        await asyncio.sleep(3)
        result = await PhindAi.create_async_generator(model, messages)  # fresh nonce
    else:
        raise

When it happens

Trigger: POSTing to admin-ajax.php with a stale/incorrect nonce or when the backend throttles: response JSON like {'success': false, 'data': 'nonce verification failed'} or a rate-limit message.

Common situations: Reusing a nonce across sessions (they are short-lived); too many requests per IP per window; long/blocked prompt content rejected server-side; g4f nonce extraction racing with nonce rotation.

Related errors

• Failed to extract nonce from PhindAi response
• Failed to decode JSON from PhindAi response: {text}
• Failed to chat: {response.status} {error_text}
• No response
• {data['code']}:{data['details']}

AI-assisted analysis of xtekky/gpt4free@973504e177 (2026-08-14). Data as JSON: /api/errors/ca210a215c496831. Report an issue: GitHub.