HeyPuter/puter · error · HttpError

internal_error

internal_error

Error message

ElevenLabs API key not configured

What it means

The call reached the ElevenLabs code path but this.#apiKey is null: config.providers.elevenlabs had no apiKey (nor api_key/key) when onServerStart ran at boot. The driver throws 500 / internal_error because a missing server-side credential is an operator defect, not a caller mistake. It fires AFTER the test_mode short-circuit, so a test call still succeeds without a key.

Source

Thrown at src/backend/drivers/ai-speech2speech/VoiceChangerDriver.ts:133

            !PROVIDERS.includes(
                args.provider
                    .trim()
                    .toLowerCase() as (typeof PROVIDERS)[number],
            )
        ) {
            throw new HttpError(
                400,
                `Speech-to-speech provider not found: ${args.provider}. Available: ${PROVIDERS.join(', ')}`,
                { legacyCode: 'bad_request' },
            );
        }

        if (args.test_mode) {
            return { url: SAMPLE_AUDIO_URL, content_type: 'audio/mpeg' };
        }

        if (!this.#apiKey) {
            throw new HttpError(500, 'ElevenLabs API key not configured', {
                legacyCode: 'internal_error',
            });
        }

        const actor = Context.get('actor');
        if (!actor)
            throw new HttpError(401, 'Authentication required', {
                legacyCode: 'unauthorized',
            });

        if (!args.audio) {
            throw new HttpError(400, '`audio` is required', {
                legacyCode: 'bad_request',
            });
        }

        const loaded = await loadFileInput(
            this.stores,

View on GitHub (pinned to 908ec23eda)

Solutions

1. Set providers.elevenlabs.apiKey (or api_key / key) in the backend config and restart so onServerStart re-reads it.
2. Verify the config file the running process actually loads (not a stale transpiled build).
3. Until configured, call with test_mode: true to return the sample audio URL, or disable the feature in the UI.

Example fix

// config.json
// before
"providers": { "elevenlabs": {} }
// after
"providers": { "elevenlabs": { "apiKey": "xi-..." } }

// probe the path without spending credits
await driver.convert({ ...args, test_mode: true });
// if the test call returns sample audio but a real call 500s with
// 'ElevenLabs API key not configured', the key is missing at the server

try {
  await driver.convert(args);
} catch (e) {
  if (e?.legacyCode === 'internal_error' && /API key not configured/.test(e?.message ?? '')) {
    showFeatureUnavailable('speech-to-speech'); // alert ops
  } else throw e;
}

When it happens

Trigger: Any real (test_mode falsy) convert() on a deployment whose config.providers.elevenlabs block is absent or has no key field, or whose key resolved to an empty/falsy string.

Common situations: Self-hosted Puter without ElevenLabs credentials in config.json; the env var never wired into the config loader; key field named with an unexpected casing the lookup (apiKey/api_key/key) does not catch.

Related errors

• ElevenLabs returned ${response.status}
• internal_error
• internal_error
• bad_request
• password_mismatch

AI-assisted analysis of HeyPuter/puter@908ec23eda (2026-08-12). Data as JSON: /api/errors/6836233277aeef51. Report an issue: GitHub.