vercel/ai · error · InvalidArgumentError

Invalid argument for parameter schemaDescription: Schema des

Error message

Invalid argument for parameter schemaDescription: Schema description is not supported for enum output.

What it means

schemaDescription annotates a schema, but output: 'enum' uses enumValues rather than a schema, so there is nothing to describe. The library throws InvalidArgumentError for schemaDescription when output is 'enum'.

Source

Thrown at packages/ai/src/generate-object/validate-object-generation-input.ts:111

      throw new InvalidArgumentError({
        parameter: 'enumValues',
        value: enumValues,
        message: 'Enum values are not supported for array output.',
      });
    }
  }

  if (output === 'enum') {
    if (schema != null) {
      throw new InvalidArgumentError({
        parameter: 'schema',
        value: schema,
        message: 'Schema is not supported for enum output.',
      });
    }

    if (schemaDescription != null) {
      throw new InvalidArgumentError({
        parameter: 'schemaDescription',
        value: schemaDescription,
        message: 'Schema description is not supported for enum output.',
      });
    }

    if (schemaName != null) {
      throw new InvalidArgumentError({
        parameter: 'schemaName',
        value: schemaName,
        message: 'Schema name is not supported for enum output.',
      });
    }

    if (enumValues == null) {
      throw new InvalidArgumentError({
        parameter: 'enumValues',
        value: enumValues,

View on GitHub (pinned to 69428b1f8b)

Solutions

Remove schemaDescription when using output: 'enum'.
Explain the enum values in the prompt if additional guidance is needed.
Attach schemaDescription only for object/array output where a schema is used.

Example fix

// before
generateObject({ model, output: 'enum', enumValues: ['a','b'], schemaDescription: 'Pick a letter', prompt });
// after
generateObject({ model, output: 'enum', enumValues: ['a','b'], prompt: 'Pick a letter. ' + prompt });

Defensive patterns

Strategy: validation

Validate before calling

if (opts.output === 'enum' && opts.schemaDescription != null) {
  throw new Error("schemaDescription must be omitted when output is 'enum'");
}

Type guard

function isEnumOptionsClean(opts) {
  return opts.output !== 'enum'
    || (opts.schema == null
      && opts.schemaDescription == null
      && opts.schemaName == null);
}

Try / catch

try {
  return await generateObject(opts);
} catch (error) {
  if (InvalidArgumentError.isInstance(error) && error.parameter === 'schemaDescription') {
    const { schemaDescription, ...rest } = opts;
    return generateObject(rest);
  }
  throw error;
}

Prevention

Only attach schemaDescription alongside an actual schema (object/array output).
Put extra guidance for enum picking into the prompt.
Sanitize options before calling generateObject/streamObject in generic wrappers.

When it happens

Trigger: Calling generateObject({ model, output: 'enum', enumValues: [...], schemaDescription: '...', prompt }) or streamObject with the same; any non-null schemaDescription with enum output triggers it.

Common situations: Generic call builder that always sets schemaDescription; converting an object-output call to enum output without pruning schema-related options; configuration driven from user input that carries stale options.

Related errors

AI-assisted analysis of vercel/ai@69428b1f8b (2026-08-30). Data as JSON: /api/errors/c91c7346cc388462. Report an issue: GitHub.