Partial Stream
Parse Partial JSON Stream - Turn your slow AI app into an engaging real-time app
FreeAIUI LibraryOpen Source10,095 impressions#3 of its week2 comments
Comments
>log in to comment
Partial Stream is a JavaScript/TypeScript library that parses partial JSON streams from OpenAI function calls in real time.
- for
- Developers building LLM-powered apps that need live streaming data handling.
- pricing
- open source
- license
- MIT
Key features
- Token-to-JSON conversion — Transforms incoming token streams into structured JSON objects as they arrive.
- Multiple UI modes — Supports All-Together, One-By-One, Progressive, and Real-time rendering styles.
- OpenAI Function Calling support — Integrates with OpenAI function calls to process partial responses early.
- Server-Sent Events output — Streams parsed entities directly to the browser via SSE for instant UI updates.
- TypeScript typings — Provides type definitions for safer integration in TS projects.
- NPM distribution — Installable via npm with a single command.
Use cases
- Display color palettes in a UI as soon as each color is generated
- Show incremental search results from an LLM without waiting for full response
- Build chat interfaces that render assistant messages token-by-token
- Create dashboards that update charts as data streams in
Partial Stream FAQ
How do I install Partial Stream?+
Run npm i openai-partial-stream and import it in your JavaScript/TypeScript code.
What modes does the library support?+
All-Together, One-By-One, Progressive (key-value pairs), and Real-time (token-by-token).
Can it work with OpenAI function calling?+
Yes, it parses the partial JSON returned by OpenAI function calls on the fly.
How are parsed entities delivered to the browser?+
The library can emit them via Server-Sent Events for immediate UI updates.
Is there documentation?+
Documentation is available in the GitHub repository linked from the website.
Summarized by DevHunt from partial.stream · Oct 2, 2026. Details may change; check the official site.
.png?width=64)
