https://github.com/joelhooks/inngest-partykit-nextjs-openai
Collaborative AI Workflows with Inngest and Partykit
https://github.com/joelhooks/inngest-partykit-nextjs-openai
Last synced: 3 months ago
JSON representation
Collaborative AI Workflows with Inngest and Partykit
- Host: GitHub
- URL: https://github.com/joelhooks/inngest-partykit-nextjs-openai
- Owner: joelhooks
- License: mit
- Created: 2023-08-24T18:25:30.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2023-09-02T23:57:16.000Z (almost 2 years ago)
- Last Synced: 2025-03-08T16:35:51.100Z (4 months ago)
- Language: TypeScript
- Homepage:
- Size: 341 KB
- Stars: 52
- Watchers: 2
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Collaborative AI-Assisted Workflows with Inngest and Partykit
This is an example using [Inngest](https://inngest.com) and
[Partykit](https://github.com/partykit/partykit) to build a collaborative
AI-assisted workflow.
📼 [quick loom video](https://www.loom.com/share/c43aa34205854096bcec0a96e7ba5634?sid=839b1adc-ad39-4540-9995-88967f2b6da9)
Inngest initiates and orchestrates the workflow, and Partykit is used
as a proxy to the OpenAI API. We need a proxy because the OpenAI API **streams**
responses, and we want to orchestrate actions based on suggestions and function
calls that the OpenAI API returns.
This is important because we aren't building a simple chat box. We want a UI that allows
us to do actual work, and we **cannot trust the LLM to make appropriate business decisions**.
We **can** trust the LLM to make suggestions, and we can use those suggestions to
make human-level decisions.
## How it works
The user has a text box they can type into and ask for a particular task. This
initiates a workflow with Inngest. As OpenAI streams back its responses, we can use Partykit
as a proxy for the streams and broadcast those responses to **every subscribed user**.
**This means that we are able to create multiplayer workflows that are assisted by AI.**
## How to use
Get everything installed:
```shell
pnpm install
```
Start everything like this:
```shell
pnpm dev
```
`pnpm dev` starts the web application server, the PartyKit dev server, and the Inngest dev server.
Inngest runs at http://localhost:8288, and the web app server runs at http://localhost:3000. Open both of these in your browser!
Partykit doesn't have a UI but is running at http://127.0.0.1:1999/
`OPENAI_API_KEY` and `LINEAR_API_KEY` need to be set in `/apps/web/.env.local` for the web app to work.
👋 Note that the linear function `delete` does NOT actually delete issues. You can create a Linear account for free to test it out.
## What's inside?
This example includes the following packages/apps:
### Apps and Packages
- `web`: simple [Next.js](https://nextjs.org/) app
- `openai-party`: a [PartyKit](https://github.com/partykit/partykit) server
- `ui`: a stub React component library shared by both `web` applications
- `eslint-config-custom`: `eslint` configurations (includes `eslint-config-next` and `eslint-config-prettier`)
- `tsconfig`: `tsconfig.json`s used throughout the monorepo
Each package/app is 100% [TypeScript](https://www.typescriptlang.org/).
### Build
To build all apps and packages, run the following command:
```
pnpm build
```
### Develop
To develop all apps and packages, run the following command:
```
pnpm dev
```