vercel/ai · error

Unsupported role: ${_exhaustiveCheck}

Error message

Unsupported role: ${_exhaustiveCheck}

What it means

convertToMistralChatMessages exhaustively switches over prompt message roles (system/user/assistant/tool). The default branch uses TypeScript's never-exhaustiveness check and throws for any role outside those, so this only fires with a role the converter doesn't know.

Source

Thrown at packages/mistral/src/convert-to-mistral-chat-messages.ts:193

            case 'content':
            case 'json':
            case 'error-json':
              contentValue = JSON.stringify(output.value);
              break;
          }

          messages.push({
            role: 'tool',
            name: toolResponse.toolName,
            tool_call_id: toolResponse.toolCallId,
            content: contentValue,
          });
        }
        break;
      }
      default: {
        const _exhaustiveCheck: never = role;
        throw new Error(`Unsupported role: ${_exhaustiveCheck}`);
      }
    }
  }

  return messages;
}

View on GitHub (pinned to 69428b1f8b)

Solutions

1. Only use system/user/assistant/tool roles in prompts sent to Mistral.
2. Update both `ai` and `@ai-sdk/mistral` to matching latest versions.
3. Fix type-unsafe prompt construction (remove `as any` casts).
4. Sanitize roles before converting legacy chat histories.

Example fix

// before
const prompt = [{ role: 'developer', content: 'hi' }] as any;
// after
const prompt = [{ role: 'system', content: 'hi' }];

const roles = ['system', 'user', 'assistant', 'tool'] as const;
function hasValidRoles(prompt: { role: string }[]): boolean {
  return prompt.every(m => (roles as readonly string[]).includes(m.role));
}

type MessageRole = 'system' | 'user' | 'assistant' | 'tool';
function isSupportedRole(role: string): role is MessageRole {
  return ['system', 'user', 'assistant', 'tool'].includes(role);
}

try {
  await generateText({ model: mistral(modelId), prompt });
} catch (error) {
  if (error instanceof Error && error.message.startsWith('Unsupported role:')) {
    console.error('Invalid role:', error.message);
    // fix prompt construction / versions, then retry
  } else {
    throw error;
  }
}

When it happens

Trigger: A LanguageModelV4Prompt containing a message role other than 'system', 'user', 'assistant', or 'tool' — typically only possible with type-unsafe prompt construction or a version mismatch between prompt types and the Mistral package.

Common situations: Casting untyped/legacy message arrays into prompts; building prompts programmatically with wrong role strings; running mismatched versions of `ai` core and `@ai-sdk/mistral`.

Related errors

• file part data type ${part.data.type}
• file parts with provider references
• text file parts
• Unsupported content type in assistant message: ${part.type}
• Only images and PDF file parts are supported

AI-assisted analysis of vercel/ai@69428b1f8b (2026-08-30). Data as JSON: /api/errors/66e777badd4daf38. Report an issue: GitHub.