BerriAI/litellm · error · NLPCloudError

{raw_response.text}

Error message

{raw_response.text}

What it means

Raised by litellm's NLPCloud chat transformer when raw_response.json() throws: nlp_cloud returned a body that is not JSON (typically an HTML error page or empty body). The message is the raw response text and the upstream status code is preserved on the NLPCloudError.

Source

Thrown at litellm/llms/nlp_cloud/chat/transformation.py:192

        optional_params: dict,
        litellm_params: dict,
        encoding: Any,
        api_key: str | None = None,
        json_mode: bool | None = None,
    ) -> ModelResponse:
        ## LOGGING
        logging_obj.post_call(
            input=None,
            api_key=api_key,
            original_response=raw_response.text,
            additional_args={"complete_input_dict": request_data},
        )

        ## RESPONSE OBJECT
        try:
            completion_response: Final = raw_response.json()
        except Exception:
            raise NLPCloudError(message=raw_response.text, status_code=raw_response.status_code)
        if "error" in completion_response:
            raise NLPCloudError(
                message=completion_response["error"],
                status_code=raw_response.status_code,
            )
        else:
            try:
                if len(completion_response["generated_text"]) > 0:
                    model_response.choices[0].message.content = completion_response["generated_text"]
            except Exception:
                raise NLPCloudError(
                    message=json.dumps(completion_response),
                    status_code=raw_response.status_code,
                )

        ## CALCULATING USAGE - baseten charges on time, not tokens - have some mapping of cost here.
        prompt_tokens: Final = completion_response["nb_input_tokens"]
        completion_tokens: Final = completion_response["nb_generated_tokens"]

View on GitHub (pinned to 6c2dcb801b)

Solutions

1. Read the exception message — it contains the literal upstream body revealing the real problem.
2. Verify NLP_CLOUD_API_KEY is set and valid.
3. If rate limited, add throttling/backoff on your side or upgrade the NLP Cloud plan.
4. Retry transient 502/503 gateway errors with exponential backoff.

Example fix

# before
out = litellm.completion(model="nlp_cloud/finetuned-llama-3-70b", messages=msgs)

# after — retry transient gateway failures
from tenacity import retry, wait_exponential, stop_after_attempt
@retry(wait=wait_exponential(multiplier=1, max=10), stop=stop_after_attempt(3), reraise=True)
def ask(msgs):
    return litellm.completion(model="nlp_cloud/finetuned-llama-3-70b", messages=msgs)
out = ask(msgs)

import os
assert os.getenv("NLP_CLOUD_API_KEY"), "NLP_CLOUD_API_KEY not set"

from litellm.exceptions import APIError
import time
for attempt in range(3):
    try:
        out = litellm.completion(model="nlp_cloud/finetuned-llama-3-70b", messages=msgs)
        break
    except APIError as e:
        status = getattr(e, "status_code", None)
        if status and status >= 500 and attempt < 2:
            time.sleep(2 ** attempt)
            continue
        raise

When it happens

Trigger: Calling litellm.completion() with model='nlp_cloud/<engine>' when the NLP Cloud gateway returns HTML (rate limit page, 502) or a plain-text error because of an invalid NLP_CLOUD_API_KEY.

Common situations: Free-tier rate limits returning an HTML block page, expired API key, or an outage of nlpcloud.io infrastructure.

Related errors

• {completion_response["error"]}
• {json.dumps(completion_response)}
• NLPCloudException - {error_str}
• An unknown error occurred with the stream
• Unable to get json response - {e}, Original Response: {raw_r

AI-assisted analysis of BerriAI/litellm@6c2dcb801b (2026-08-15). Data as JSON: /api/errors/b1e5b3976f3f664c. Report an issue: GitHub.