Ink 2 is our latest streaming speech-to-text model. It’s super fast with the best turn-detection and lowest error rate of any streaming model — engineered for enterprise voice agents. To learn how to use Ink 2, see the Speech-to-Text API reference.Documentation Index
Fetch the complete documentation index at: https://docs.cartesia.ai/llms.txt
Use this file to discover all available pages before exploring further.
Model snapshots
| Snapshot | Release Date | Languages | Status |
|---|---|---|---|
ink-2 | May 22, 2026 | en | Preview |
- the latest stable snapshot of the model
- the preview snapshot
Continuous updates
Using the modelink-2 in your API calls will automatically keep you up to date with the most recent stable snapshot of the model (or preview snapshot if no stable snapshots exist).
If pinning a specific version is important for your use case, we recommend switching to the dated version once it becomes available.
| Model ID | Model update behavior | Recommended for |
|---|---|---|
ink-2-YYYY-MM-DD | Snapshotted, will never change | Customers who want to run internal evals before any updates |
ink-2 | Will be updated to point to the most recent stable snapshot | Customers who want stable releases, but want to be up-to-date with the recent capabilities |
Older Models
For information onink-whisper, see our page on Older STT Models.