Installation

TypeScript browser packages

Most browser apps start with @slop-ai/client, then add the framework adapter that matches the UI layer.

bun add @slop-ai/client

Framework	Package	Install
React	@slop-ai/react	bun add @slop-ai/client @slop-ai/react
Vue 3	@slop-ai/vue	bun add @slop-ai/client @slop-ai/vue
SolidJS	@slop-ai/solid	bun add @slop-ai/client @slop-ai/solid
Angular	@slop-ai/angular	bun add @slop-ai/client @slop-ai/angular
Svelte 5	@slop-ai/svelte	bun add @slop-ai/client @slop-ai/svelte
Vanilla JS / TS	@slop-ai/client	bun add @slop-ai/client

Use @slop-ai/core when you need the shared types, helpers, and tree utilities directly:

bun add @slop-ai/core

TypeScript server and consumer packages

Package	Use case	Install
@slop-ai/server	Node.js, Bun, desktop helpers, and CLI providers	bun add @slop-ai/server
@slop-ai/consumer	custom agents, inspectors, bridges, and tests	bun add @slop-ai/consumer
@slop-ai/tanstack-start	TanStack Start full-stack integration	bun add @slop-ai/server @slop-ai/tanstack-start
@slop-ai/openclaw-plugin	OpenClaw integration	bun add @slop-ai/openclaw-plugin

Python

pip install slop-ai[websocket]

from fastapi import FastAPI
from slop_ai import SlopServer
from slop_ai.transports.asgi import SlopMiddleware

app = FastAPI()
slop = SlopServer("my-api", "My API")
app.add_middleware(SlopMiddleware, slop=slop)

See the Python guide and Python API page for transport and consumer examples.

Go

go get github.com/devteapot/slop/packages/go/slop-ai

package main

import (
  "net/http"

  slop "github.com/devteapot/slop/packages/go/slop-ai"
)

func main() {
  server := slop.NewServer("my-app", "My App")
  server.Mount(http.DefaultServeMux)
  http.ListenAndServe(":8080", nil)
}

See the Go guide and Go API page for Unix, stdio, and consumer support.

Rust

cargo add slop-ai

use serde_json::json;
use slop_ai::SlopServer;

let slop = SlopServer::new("my-app", "My App");
slop.register("status", json!({"type": "status", "props": {"healthy": true}}));

See the Rust guide and Rust API page for feature-flag and transport details.

Apps

The shipped consumer apps both include AI chat, but chat only works once you have an active model profile configured.

Both apps start with a default local profile:

• provider: Ollama
• endpoint: http://localhost:11434
• model: qwen2.5:14b

If you are using OpenAI, OpenRouter, or Gemini instead, add your endpoint and API key before chatting.

Chrome extension

cd apps/extension
bun install
bun run build

Load the apps/extension directory in chrome://extensions as an unpacked extension.

After loading the extension:

1. Open the popup.
2. Use LLM Settings -> to open the options page.
3. Add or edit a profile for Ollama, OpenAI, OpenRouter, or Gemini.
4. Open the in-page chat overlay and pick the active profile and model.

Desktop app

cd apps/desktop
bun install
bun run dev

The Tauri app will build the frontend and launch the native desktop shell.

After launching the desktop app:

1. Open Settings.
2. Add or edit a profile for Ollama, OpenAI, OpenRouter, or Gemini.
3. Use the top-bar selectors to switch the active profile and model.