# Replicate — Streaming output for language models

- Company: Replicate (replicate.com)
- Announced: 2023-08-14
- Category: capability-change
- Coverage: not counted
- Announcement: yes
- Group: announcements
- Source: https://replicate.com/blog/streaming
- Record: https://forck.live/items/2559-streaming-output-for-language-models
- Subject: Platform
- Models affected: llama-2-70b-chat, Falcon, Vicuna, StableLM, Llama 2

Replicate's API now supports server-sent event streams for language models, allowing users to receive real-time token-by-token output by setting the stream option to true and connecting to a stream URL.

## Evidence

Verbatim from https://replicate.com/blog/streaming:

> Replicate's API now supports server-sent event streams for language models.

---

Record: https://forck.live/items/2559-streaming-output-for-language-models
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md