OpenRouter adds ElevenLabs speech models, with launch pricing half off

OpenRouter has added nine ElevenLabs text-to-speech models and two speech-to-text models, letting developers add voice input and output through its API. All ElevenLabs models on OpenRouter are listed at 50% off until 19 October at 8am PT, according to the company’s launch guide.

OpenRouter Watch analysis

What happened

The new options include Eleven v4 for expressive narration, v4 Turbo for lower-latency voice-agent replies, and Scribe v2 for transcription with speaker labels and word timestamps. Developers can call OpenRouter’s speech and transcription endpoints using an OpenRouter API key, without a separate ElevenLabs plan.

The guide also lays out practical limits. Text-to-speech requests have model-specific character caps, from 5,000 to 40,000 characters. Scribe uploads are capped at 25 MB, and very long audio can run into an upstream request timeout. Realtime Scribe over WebSocket is not part of this launch. The discount applies through the stated deadline; the guide points users to each model’s page for current per-character prices.

Why it matters

For developers already routing language-model requests through OpenRouter, speech can now sit alongside them behind the same API access. That makes a voice-agent prototype or a transcription step simpler to wire up, while leaving the choice of chat model separate from the audio models.

The immediate bargain is real but temporary. The useful comparison is not simply “one key, fewer chores”: teams should check character or audio limits, model fit and the price after the promotion before building a service around it. A tidy API shortcut is welcome; a surprise bill is a less charming feature.

Our read

This is a substantial addition to OpenRouter’s model catalogue, with concrete uses ranging from narrated content to diarised meeting notes. Start with a small test: try the voice and transcript quality on representative material, check request limits, and work out the post-discount cost before committing. The guide describes the available features, not an independent quality comparison.

What to watch

• Whether the promotion ends as stated on 19 October at 8am PT.
• How prices and usage limits compare with using ElevenLabs directly.
• Whether realtime transcription becomes available through OpenRouter.

Discussion spark: For a voice-enabled app, would you rather keep speech and language models behind one API, or use separate providers if that gives you clearer pricing and control?

not affiliated with or endorsed by OpenRouter