<br />

Firebase AI Logicsupports basic**streaming of text responses** using`generateContentStream`or`sendMessageStream`(for chat).

You can achieve faster interactions by not waiting for the entire result from the model generation, and instead use streaming to handle partial results.

Check out the documentation to stream generated text responses for the following:

- [Text-only input](https://firebase.google.com/docs/ai-logic/generate-text#streaming)

- [Analyzing images](https://firebase.google.com/docs/ai-logic/analyze-images#streaming)

- [Analyzing video](https://firebase.google.com/docs/ai-logic/analyze-video#streaming)

- [Analyzing audio](https://firebase.google.com/docs/ai-logic/analyze-audio#streaming)

- [Analyzing documents (like PDFs)](https://firebase.google.com/docs/ai-logic/analyze-documents#streaming)

- [Chat experiences](https://firebase.google.com/docs/ai-logic/chat#streaming)

| **Important:** If you're looking for low-latency*live bidirectional streaming* of text and voice (for example, to let your users experience natural, human-like voice conversations), then check out[Bidirectional streaming using theGemini Live API](https://firebase.google.com/docs/ai-logic/live-api).