| name | openui-forge-go |
| description | OpenUI generative UI with Go (net/http) backend. Direct OpenAI API streaming via HTTP. |
| version | 1.2.0 |
| author | OthmanAdi |
OpenUI Forge — Go
Build generative UI apps with a React frontend + Go backend. Streams OpenAI API responses directly via net/http.
Activation Triggers
- "openui go", "openui golang", "openui go backend"
- "generative ui go", "go streaming ui backend"
Prerequisites
- Node.js >= 22 (24 LTS recommended) + React >= 18.3.1 (19+ recommended) (frontend)
- Go >= 1.24 (backend; 1.23 and older are out of security support as of Go 1.26)
OPENAI_API_KEY environment variable set
Quick Start
- Create the React frontend and install OpenUI deps:
npm install @openuidev/react-ui @openuidev/react-headless @openuidev/react-lang lucide-react zod
- Generate the system prompt:
npx @openuidev/cli generate ./src/lib/library.ts --out backend/system-prompt.txt
- Create the Go backend (see Full Code below)
- Run:
go run main.go on :8080, frontend on :3000
Full Code
Backend: backend/go.mod
module openui-backend
go 1.24
require (
github.com/joho/godotenv v1.5.1
)
Backend: backend/main.go
package main
import (
"bufio"
"bytes"
"encoding/json"
"fmt"
"log"
"net/http"
"os"
_ "github.com/joho/godotenv/autoload"
)
var systemPrompt string
func init() {
data, err := os.ReadFile("system-prompt.txt")
if err != nil {
log.Fatal("system-prompt.txt not found: ", err)
}
systemPrompt = string(data)
}
func corsMiddleware(next http.Handler) http.Handler {
return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
w.Header().Set("Access-Control-Allow-Origin", "http://localhost:3000")
w.Header().Set("Vary", "Origin")
w.Header().Set("Access-Control-Allow-Methods", "POST, OPTIONS")
w.Header().Set("Access-Control-Allow-Headers", "Content-Type, Authorization")
if r.Method == "OPTIONS" {
w.WriteHeader(204)
return
}
next.ServeHTTP(w, r)
})
}
type Message struct {
Role string `json:"role"`
Content string `json:"content"`
}
type ChatRequest struct {
Messages []Message
}
{
r.Method != {
http.Error(w, , )
}
req ChatRequest
err := json.NewDecoder(r.Body).Decode(&req); err != {
http.Error(w, , )
}
messages := ([]Message{{Role: , Content: systemPrompt}}, req.Messages...)
model := os.Getenv()
model == {
model =
}
body, _ := json.Marshal([]{}{
: model, : , : messages,
})
apiReq, _ := http.NewRequest(, , bytes.NewReader(body))
apiReq.Header.Set(, )
apiReq.Header.Set(, +os.Getenv())
resp, err := http.DefaultClient.Do(apiReq)
err != {
http.Error(w, , )
}
resp.Body.Close()
w.Header().Set(, )
w.Header().Set(, )
w.Header().Set(, )
flusher, ok := w.(http.Flusher)
!ok {
http.Error(w, , )
}
scanner := bufio.NewScanner(resp.Body)
scanner.Buffer(([], , *), *)
scanner.Scan() {
line := scanner.Bytes()
_, err := w.Write(line); err != {
}
_, err := w.Write([]()); err != {
}
flusher.Flush()
}
}
{
mux := http.NewServeMux()
mux.HandleFunc(, chatHandler)
fmt.Println()
log.Fatal(http.ListenAndServe(, corsMiddleware(mux)))
}
Frontend: app/chat/page.tsx
"use client";
import { FullScreen } from "@openuidev/react-ui";
import { openuiChatLibrary } from "@openuidev/react-ui/genui-lib";
import {
openAIAdapter,
openAIMessageFormat,
} from "@openuidev/react-headless";
export default function ChatPage() {
return (
<FullScreen
componentLibrary={openuiChatLibrary}
streamProtocol={openAIAdapter()}
messageFormat={openAIMessageFormat}
apiUrl="http://localhost:8080/api/chat"
/>
);
}
The Go backend forwards OpenAI's SSE stream line-by-line with a bufio.Scanner loop, flushing after each line (no io.Copy), so the client sees tokens as they arrive. Pair it with openAIAdapter() on the frontend. openAIReadableStreamAdapter() is for NDJSON (no data: prefix) and will silently produce no output here.
An official OpenAI Go SDK exists (github.com/openai/openai-go/v3, ~v3.41.0, requires Go 1.22+) as an alternative to hand-rolling the HTTP call. This skill keeps raw net/http as the dependency-free default. Note the SDK still ships small, limited breaking changes between releases (per its own README, some backwards-incompatible changes land in minor versions), so pin a version if you adopt it.
System Prompt Generation
npx @openuidev/cli generate ./src/lib/library.ts --out backend/system-prompt.txt
Validation Checklist
Error Patterns
| Error | Cause | Fix |
|---|
| CORS blocked | Origin mismatch | Update Access-Control-Allow-Origin in middleware |
system-prompt.txt not found | File missing from backend dir | Run CLI generate command |
| 502 Bad Gateway | OpenAI API unreachable or key invalid | Check OPENAI_API_KEY and network |
| Stream not flushing | Missing http.Flusher | Ensure handler calls flusher.Flush() |
| Empty response | Body not forwarded | Verify the bufio.Scanner loop writes and flushes each line |
