vercel/ai · error

Unsupported chunk type: ${_exhaustiveCheck}

Error message

Unsupported chunk type: ${_exhaustiveCheck}

What it means

The array-mode element stream transform exhaustively switches over every ObjectStreamPart type and throws this Error in the `default` branch. TypeScript's `_exhaustiveCheck: never` guarantees this is unreachable when the switch covers all chunk types, so at runtime it signals that the library received a stream part type it does not recognize — usually a version mismatch between the `ai` package internals or corrupted/intercepted stream data.

Source

Thrown at packages/ai/src/generate-object/output-strategy.ts:297

                  for (
                    ;
                    publishedElements < array.length;
                    publishedElements++
                  ) {
                    controller.enqueue(array[publishedElements]);
                  }

                  break;
                }

                case 'text-delta':
                case 'finish':
                case 'error': // suppress error (use onError instead)
                  break;

                default: {
                  const _exhaustiveCheck: never = chunk;
                  throw new Error(
                    `Unsupported chunk type: ${_exhaustiveCheck}`,
                  );
                }
              }
            },
          }),
        ),
      );
    },
  };
};

const enumOutputStrategy = <ENUM extends string>(
  enumValues: Array<ENUM>,
): OutputStrategy<string, ENUM, never> => {
  return {
    type: 'enum',

View on GitHub (pinned to 69428b1f8b)

Solutions

1. Run `pnpm install` freshly and verify only one version of `ai` (and `@ai-sdk/*`) is in the lockfile (`pnpm why ai`).
2. Rebuild after upgrading so the transform and the stream-part types come from the same package version.
3. Remove or fix custom middleware/provider patches that inject unknown chunk types into the object stream.
4. If reproducible on one consistent version, file a bug with the chunk type shown in the message.

Example fix

// before (mixed versions)
"dependencies": { "ai": "^4.0.0", "@ai-sdk/openai": "^1.0.0" } // plus stale ai@3 in lockfile

// after
rm -rf node_modules && pnpm install
"dependencies": { "ai": "^4.3.16", "@ai-sdk/openai": "^1.3.22" }

// before streaming, assert a single SDK version at startup
import { VERSION } from 'ai';
console.assert(VERSION === expectedVersion, 'mixed ai versions detected');

function isKnownChunkType(t: string): boolean {
  return ['object','text-delta','finish','error','start','start-step','finish-step','reasoning-delta','source','tool-call','tool-input-delta','tool-result','raw'].includes(t);
}

try {
  for await (const el of result.elementStream) { /* ... */ }
} catch (e) {
  if (e instanceof Error && e.message.startsWith('Unsupported chunk type')) {
    console.error('SDK version skew; reinstall deps', e.message);
    // fallback: consume full object instead of element stream
  } else throw e;
}

When it happens

Trigger: A new ObjectStreamPart variant is emitted by a newer/patched `ai` runtime while this transform is from a different version (mixed package versions, duplicated `ai` installs, bundler deduplication issues). Also possible when stream chunks are injected/modified by custom middleware or patched providers.

Common situations: Monorepos with two versions of `ai` resolved simultaneously; stale node_modules after an upgrade; custom provider wrappers that emit non-standard stream parts into the object stream.

Related errors

• Unsupported chunk type: ${_exhaustiveCheck}
• Unknown chunk type: ${exhaustiveCheck}
• 'element streams in no-schema mode' functionality not suppor
• 'element streams in object mode' functionality not supported
• 'element streams in enum mode' functionality not supported.

AI-assisted analysis of vercel/ai@69428b1f8b (2026-08-30). Data as JSON: /api/errors/e2dbc130c4cd7271. Report an issue: GitHub.