Setting up narration voices

Which voice should I pick?

• Google Cloud TTS recommended — natural voices, a generous free tier (1 million characters per month), and exact word-by-word highlighting while you listen. Best all-round choice.
• ElevenLabs — the most expressive voices; smaller free tier. Also supports exact word highlighting.
• Gemini — very natural, expressive voices, but word highlighting is approximate (estimated, not exact) and generation is slower.
• xAI — Grok voices; word highlighting is approximate.
• Amazon Polly — the broadest German coverage (Germany, Austria, Switzerland), cheap Standard voices and higher-quality Neural voices, and exact word highlighting (Speech Marks). Needs AWS keys and a region.
• Azure AI Speech — a very large German neural catalog and a recurring free tier; word highlighting is approximate (the plain REST voice API returns no word timing).

Tip: if you mainly want to read along word-for-word, use Google Cloud, ElevenLabs, or Amazon Polly.

Google Cloud TTS

You need a Google account and a Google Cloud project with billing enabled. The free tier covers 1 million characters a month, so normal reading stays free — but a card on file is required by Google.

1. Go to console.cloud.google.com and create or pick a project.
2. Open APIs & Services → Library, search Cloud Text-to-Speech API, and click Enable.
3. If asked, link a billing account (Billing in the left menu).
4. Open APIs & Services → Credentials → Create credentials → API key.
5. (Recommended) Click Edit API key → Restrict key and limit it to Cloud Text-to-Speech API.
6. Copy the key, then in Voice Compiler open Settings → Provider API keys, select Google Cloud, paste, Save key, then Test.

Note: this is not the same key as Gemini. A Cloud Text-to-Speech key does not work for Gemini, and vice versa.

Gemini

The quickest key to get — no billing required for the free tier. It comes from Google AI Studio, which is separate from Google Cloud above.

1. Go to aistudio.google.com/apikey and sign in with your Google account.
2. Click Create API key (you can create it in a new or existing project).
3. Copy the key, then in Voice Compiler open Settings → Provider API keys, select Gemini, paste, Save key, then Test.

Heads-up: Gemini voices sound great, but the read-along highlight is estimated, so a word can drift slightly from the audio. For exact highlighting use Google Cloud or ElevenLabs.

ElevenLabs

1. Create a free account at elevenlabs.io.
2. Open your profile menu and choose API Keys (or visit Settings → API Keys).
3. Click Create API Key and copy it.
4. In Voice Compiler open Settings → Provider API keys, select ElevenLabs, paste, Save key, then Test.

The free tier has a monthly character limit; long articles use it up quickly. Voice Compiler lists the voices of your ElevenLabs account. ElevenLabs' built-in Default voices exist only for accounts created before March 2026 and stop working on 31 December 2026.

xAI (Grok)

1. Sign in at console.x.ai.
2. Open API Keys and create a new key (this provider is paid — add credit first).
3. In Voice Compiler open Settings → Provider API keys, select xAI, paste, Save key, then Test.

Inworld

Inworld is a low-cost cloud TTS. German voices appear from its live catalog once a key is set. Word highlighting is approximate — Voice Compiler estimates timing (turn on neural CTC alignment for more accuracy); Inworld's own high-accuracy timestamps are English-only.

1. Create an account at inworld.ai and open the developer portal (the free tier includes a one-time evaluation allowance).
2. Open API Keys, generate a key, and copy its Base64 form — Voice Compiler expects that ready-made Basic string, so no extra encoding is needed.
3. In Voice Compiler open Settings → Provider API keys, select Inworld, paste, Save key, then Test.

Models: TTS-2 is Inworld's flagship with the best quality; TTS-2 Flash is the fastest and cheapest, with the same languages. The older 1.5 Max and 1.5 Mini are deprecated and only kept so earlier narrations still work.

Speechify

Speechify returns word-level timestamps with every request, so read-along is word-exact without the estimated fallback and without a second billed call. Mind the model: Simba 3.2 is the flagship but is English only — pick Simba 3.0 for German.

1. Sign up at speechify.ai and open the API console. The free plan gives 500,000 characters per month, recurring, with no card required.
2. Create an API key (it starts with sk_) and copy it.
3. In Voice Compiler open Settings → Provider API keys, select Speechify, paste, Save key, then Test.

Beyond the free allowance, paid use starts at the Starter plan: $10 per month including 1.9 million characters, then $10 per additional million. Unlike some providers there is no pay-as-you-go tier without a monthly plan. Speechify's terms also ask that audio played to other people is disclosed as AI-generated.

Amazon Polly

Polly uses AWS credentials: an access key ID, a secret access key, and a region. You paste all three in Settings → Provider API keys (select Amazon Polly). Polly gives exact word-by-word highlighting and the widest German voice range (de-DE, de-AT, de-CH).

1. Sign in to the AWS Console (or create a free account — Polly's free tier covers the first 12 months).
2. Open IAM → Users → Create user, then attach the AmazonPollyReadOnlyAccess policy (it allows DescribeVoices and SynthesizeSpeech — nothing else is needed).
3. On that user open Security credentials → Create access key and copy both the Access key ID and the Secret access key (the secret is shown once).
4. Pick a region close to you that supports Polly, e.g. eu-central-1 (Frankfurt) or us-east-1.
5. In Voice Compiler open Settings → Provider API keys, select Amazon Polly, paste the secret access key in API key, fill in Access key ID and AWS region, Save key, then Test.

Pricing: Standard voices are the cheapest but sound more robotic; Neural voices cost more but sound natural. Speech Marks (word timing) are billed at the same rate as the audio — no extra charge.

Azure AI Speech

Azure needs a subscription key and the region of your Speech resource, both pasted in Settings → Provider API keys (select Azure AI Speech). Azure has a very large German neural catalog and a recurring free tier (F0). Note: word highlighting is approximate — the plain voice REST API returns no word timing, so Voice Compiler estimates it (turn on neural CTC alignment for more accuracy).

1. Sign in to the Azure Portal and Create a resource → Speech (the free F0 tier gives 500,000 characters a month).
2. Open the created Speech resource, then Keys and Endpoint.
3. Copy KEY 1 and note the Location/Region shown there (e.g. westeurope).
4. In Voice Compiler open Settings → Provider API keys, select Azure AI Speech, paste the key in API key, fill in Azure region, Save key, then Test.

The model picker also offers MAI-Voice-2 and the faster, cheaper MAI-Voice-2-Flash (both preview) — Microsoft's newer, more expressive voices, German included. They use the same key but are not covered by the free tier, and a resource in the germanywestcentral or northeurope region does not offer them. When you pick an Azure Neural voice, the narration dialog shows how many of your 500,000 free monthly characters are left and warns before a generation would exceed them.

Custom pricing (optional)

Each provider row in Settings → Provider API keys has an optional Custom rate (USD per 1M characters) field. Set it only if your plan or discount differs from the list price — for example an ElevenLabs subscription or an AWS volume rate. Leave it empty to use the provider's list/API price. This changes the in-app cost estimate and the cost ledger only; it never affects what the provider actually bills you.

What does “Test” do?

Test asks the provider to speak a single word using your key. A green ✓ Connection OK means the key works. A red message usually means the key is wrong, billing isn’t enabled, or the provider is out of quota. The badge next to each provider shows whether a key is currently stored.

Accounts and signing in

Your library lives on the server in an account. There are three ways in:

• Google — sign in with your Google account.
• Username and password — no email address needed. Use at least 10 characters; a passphrase of a few unrelated words is best. When you create the account you get a recovery code: keep it somewhere safe, such as a password manager. It is the only way back in after a forgotten password, because there is no email reset. Using it sets a new password, gives you a new code and signs out your other devices.
• Try without an account — a guest account on the server with every feature, up to 100 MiB and articles of up to 200,000 characters. It is deleted with everything in it after 30 days without a visit. These three limits are defaults that the server's operator can change; the app states the period that applies. Add a password or link Google to keep it; signing out of a guest account that has neither leaves its library unreachable.

Settings → Login methods links Google to a password account and the other way round, changes the password, creates a new recovery code, and removes a login method as long as another one remains. If you link a Google account that already has its own library while you are a guest, the app offers to move your guest articles into it.

After too many wrong passwords, sign-in pauses for a while and the message says how long. If your browser blocks cookies for this site, the app keeps your session as a token in this browser's storage instead and says so when you sign in.

Articles from the old browser-only mode: if this browser still has them, Settings offers Move into this account. A JSON backup exported from that mode can be restored with Settings → Import library.

Keyboard shortcuts

Whenever a narration is loaded, these keys control playback from any page. They are ignored while typing in a text field or with a dialog open, and a focused control keeps the keys it uses itself: Space and Enter on a focused button press that button, and the arrow keys move within a list of choices.

• Space or K — play / pause
• ← / → or J / L — skip back / forward by your configured seek amount (Settings → Playback)
• H — jump back to the start of the article
• + / - — speed up / slow down in 0.25× steps
• ? — show this list inside the app

The same keys drive device Read aloud, where skipping moves by paragraph instead of by seconds.

If you use speech input or another tool that types single letters, switch off Settings → Keyboard shortcuts → Single-key shortcuts: the letter, symbol and Space shortcuts then do nothing, and the arrow keys keep working.