langgenius/dify · warning · ConversationCompletedError

conversation_completed

conversation_completed

Error message

The conversation has ended. Please start a new conversation.

What it means

Raised by CompletionApi.post (HTTP 400, error_code 'conversation_completed') when AppGenerateService.generate raises services.errors.conversation.ConversationCompletedError. The targeted conversation has been explicitly ended/completed and will no longer accept new messages. This is a domain-level lifecycle guard from the conversation service, re-thrown here as an HTTP error.

Source

Thrown at api/controllers/console/explore/completion.py:129

        installed_app.last_used_at = naive_utc_now()
        db.session.commit()

        try:
            response = AppGenerateService.generate(
                session=session,
                app_model=app_model,
                user=current_user,
                args=args,
                invoke_from=InvokeFrom.EXPLORE,
                streaming=streaming,
            )

            # response-contract:ignore compact_generate_response
            return helper.compact_generate_response(response)
        except services.errors.conversation.ConversationNotExistsError:
            raise NotFound("Conversation Not Exists.")
        except services.errors.conversation.ConversationCompletedError:
            raise ConversationCompletedError()
        except services.errors.app_model_config.AppModelConfigBrokenError:
            logger.exception("App model config broken.")
            raise AppUnavailableError()
        except ProviderTokenNotInitError as ex:
            raise ProviderNotInitializeError(ex.description)
        except QuotaExceededError:
            raise ProviderQuotaExceededError()
        except ModelCurrentlyNotSupportError:
            raise ProviderModelCurrentlyNotSupportError()
        except InvokeError as e:
            raise CompletionRequestError(e.description)
        except ValueError as e:
            raise e
        except Exception:
            logger.exception("internal server error.")
            raise InternalServerError()

View on GitHub (pinned to ef8544b173)

Solutions

1. Start a brand-new completion request without the completed conversation reference.
2. On the client, treat this error code as a signal to reset the conversation UI.
3. Review whether an explicit stop or a TTL policy closed the conversation.
4. If conversations should remain open, check the app's conversation lifecycle settings.

Example fix

// before: reusing a conversation handle after completion
if (err.code === 'conversation_completed') { retryWithSameConversation() }

// after: start fresh when the server says the conversation ended
if (err.code === 'conversation_completed') { startNewConversation() }

// For completion apps there is usually no conversation to check; if you carry one,
// verify it is not completed before posting. Requires a conversation-status API if available.
// Otherwise, the safest validation is simply to start a new completion without state.
if (carriedConversationId) { /* completion apps are stateless — prefer not sending it */ }

function isConversationUsable(conv) {
  return Boolean(conv && conv.status !== 'completed' && conv.status !== 'ended');
}

try {
  await postCompletion(id, payload);
} catch (err) {
  if (err.code === 'conversation_completed') {
    // reset state and start fresh; do not retry the same conversation
    startNewCompletion();
  } else { throw err; }
}

When it happens

Trigger: POST /console/installed-apps/<id>/completion-messages carrying args that reference a conversation already marked completed. The shared generate pipeline detects the conversation's completed state and aborts before invoking the model.

Common situations: Client continues to send to a conversation the server already closed (timeout, explicit stop, or completion); a long-lived client did not refresh after the conversation ended; completion app was wired to a conversation that a background job closed.

Related errors

• Conversation Not Exists.
• conversation_completed
• app_unavailable
• not_completion_app
• provider_not_initialize

AI-assisted analysis of langgenius/dify@ef8544b173 (2026-08-12). Data as JSON: /api/errors/aba14311337c6deb. Report an issue: GitHub.