mastra-ai/mastra · error · SchemaValidationError

SchemaValidationError(field, this.formatErrors(result.error)

Error message

SchemaValidationError(field, this.formatErrors(result.error))

What it means

The dataset validator converts a JSON Schema (JSONSchema7) to a Zod schema and safeParses the data. When parsing fails, it throws SchemaValidationError carrying the field name ('input' or 'groundTruth') and the formatted Zod error list. This means a dataset item does not conform to the declared schema for the given field.

Source

Thrown at packages/core/src/datasets/validation/validator.ts:41

    if (!zodSchema) {
      const zodString = jsonSchemaToZod(schema);
      zodSchema = resolveZodSchema(zodString);
      this.cache.set(cacheKey, zodSchema);
    }
    return zodSchema;
  }

  /** Clear cached validator (call when schema changes) */
  clearCache(cacheKey: string): void {
    this.cache.delete(cacheKey);
  }

  /** Validate data against schema */
  validate(data: unknown, schema: JSONSchema7, field: 'input' | 'groundTruth', cacheKey: string): void {
    const zodSchema = this.getValidator(schema, cacheKey);
    const result = zodSchema.safeParse(data);
    if (!result.success) {
      throw new SchemaValidationError(field, this.formatErrors(result.error));
    }
  }

  /** Validate multiple items, returning valid/invalid split */
  validateBatch(
    items: Array<{ input: unknown; groundTruth?: unknown }>,
    inputSchema: JSONSchema7 | null | undefined,
    outputSchema: JSONSchema7 | null | undefined,
    cacheKeyPrefix: string,
    maxErrors = 10,
  ): BatchValidationResult {
    const result: BatchValidationResult = { valid: [], invalid: [] };

    // Pre-compile schemas for performance
    const inputValidator = inputSchema ? this.getValidator(inputSchema, `${cacheKeyPrefix}:input`) : null;
    const outputValidator = outputSchema ? this.getValidator(outputSchema, `${cacheKeyPrefix}:output`) : null;

    for (const [i, item] of items.entries()) {

View on GitHub (pinned to 75dd419e61)

Solutions

1. Read the formatted error list in the SchemaValidationError and correct the offending item(s) to match the schema.
2. Run the validation locally first (e.g. with ajv) before submitting records so failures are caught client-side.
3. If the schema changed intentionally, migrate existing records to the new shape instead of editing the schema to fit old data.
4. Confirm the correct JSONSchema7 draft is used; unsupported or malformed schemas can compile to an unexpectedly strict Zod schema.

Example fix

// before
await dataset.addRecord({ input: { q: 'hello' } }); // schema requires 'question'
// after
await dataset.addRecord({ input: { question: 'hello' } });

import Ajv from 'ajv';
const ajv = new Ajv();
const validate = ajv.compile(schema);
if (!validate(item.input)) throw new Error(`Invalid input: ${ajv.errorsText(validate.errors)}`);

function conformsToSchema(data: unknown, schema: object): boolean {
  try {
    const ajv = new Ajv();
    return Boolean(ajv.compile(schema as object)(data));
  } catch { return false; }
}

try {
  dataset.validate(data, schema, 'input', cacheKey);
} catch (e) {
  if (e.name === 'SchemaValidationError') {
    console.error(`Field '${e.field}' failed:`, e.message);
    return; // or fix/reject the record
  }
  throw e;
}

When it happens

Trigger: Calling validator.validate(data, schema, field, cacheKey) directly, or any library path that validates dataset items (saving/adding records, creating experiments, running evaluation) with data that violates the JSON schema for input or groundTruth.

Common situations: Missing required properties, wrong types (e.g. string instead of number), extra strictness from the schema, or a dataset row pasted in from another source with a different shape; also happens when the schema was changed after records were created with an older shape.

Understand the failure class

Background: Schema validation failed / invalid input schema: payload rejected because its shape doesn't match the expected schema — this error's family across 28 libraries.

Related errors

• We could not convert the schema to a JSONSchema
• SchemaUpdateValidationError
• Schema validation failed due to an invalid schema definition
• WORKFLOW_SCHEMA_VALIDATION_FAILED
• ${label} contains an unsupported field.

AI-assisted analysis of mastra-ai/mastra@75dd419e61 (2026-08-30). Data as JSON: /api/errors/b31853c540eec874. Report an issue: GitHub.