openwebui-go

A complete Go client SDK for the Open WebUI REST API.
Covers all 462 endpoints across 29 domain tags — streaming, auth, Ollama & OpenAI proxies, RAG, and more.

Install • Quickstart • Configuration • API Coverage • Error Handling • Streaming • Support

---

Install

go get github.com/notfixingit3/openwebui-go

Requires Go 1.24 or later.

Quickstart

package main

import (
    "context"
    "fmt"
    "log"
    "os"

    "github.com/notfixingit3/openwebui-go/client"
)

func main() {
    apiKey := os.Getenv("OPENWEBUI_API_KEY")
    if apiKey == "" {
        log.Fatal("OPENWEBUI_API_KEY not set")
    }

    c, err := client.New(client.Config{APIKey: apiKey})
    if err != nil {
        log.Fatal(err)
    }

    // Check service health (public endpoint, no auth required)
    version, err := c.GetAppVersion(context.Background())
    if err != nil {
        log.Fatal(err)
    }
    fmt.Printf("Open WebUI version: %s\n", version["version"])
}

Configuration

client.Config accepts three fields:

Field	Default	Description
APIKey	OPENWEBUI_API_KEY env var	Bearer token for authenticated endpoints
BaseURL	http://localhost:8080	Root URL of the Open WebUI server
HTTPClient	&http.Client{Timeout: 30s}	Custom HTTP client (timeouts, TLS, proxies)

// Custom timeout and TLS config
c, err := client.New(client.Config{
    APIKey:  "sk-...",
    BaseURL: "https://openwebui.example.com",
    HTTPClient: &http.Client{
        Timeout: 60 * time.Second,
        Transport: &http.Transport{TLSClientConfig: &tls.Config{...}},
    },
})

API Coverage

The SDK covers all 462 endpoints across 29 domain tags.

Domain	Endpoints	Key Operations
analytics	8	Daily stats, messages, models, tokens, users
audio	6	TTS, STT, voices, config
auths	21	Signin, signup, OAuth, LDAP, API keys
automations	8	CRUD, toggle, runs
calendars	13	CRUD, events, RSVP
channels	28	Messaging, threads, reactions, webhooks, members
chats	42	CRUD, archive, share, clone, tag, search, stats
configs	20	Server config, banners, connections, code execution
evaluations	16	Feedback, leaderboard, config
files	12	Upload, content retrieval, processing, search
folders	7	CRUD, parent, expanded state
functions	17	Plugin functions, valves, toggle, sync
groups	10	CRUD, user membership, export
images	6	Generation, edit, config
knowledge	17	Knowledge bases, files, search, reindex, access
memories	7	CRUD, query, reset
models	14	CRUD, tags, profile images, base models, import/export
notes	9	CRUD, pin, search, access
ollama	43	Proxy: chat, generate, embeddings, model management
openai	12	OpenAI-compatible proxy, chat completions, audio
pipelines	8	Add, upload, valves
prompts	17	Prompt templates, history, versioning, tags
retrieval	16	RAG: embeddings, process file/text/web/youtube
skills	9	CRUD, access, toggle, export
tasks	11	Auto-title, emoji, tags, follow-up, MOA completions
terminals	8	Terminal server proxy
tools	15	Tool plugins, valves, access, load from URL
users	21	CRUD, groups, permissions, settings, profile images
utils	6	DB download, code execute/format, markdown, PDF

All generated request and response types live in internal/gen and are re-exported through the client package where appropriate.

Error Handling

Non-2xx responses are returned as *client.HTTPError, which exposes the status code and raw response body.

var httpErr *client.HTTPError
if errors.As(err, &httpErr) {
    fmt.Printf("HTTP %d: %s\n", httpErr.StatusCode, string(httpErr.Body))
}

Validation errors (422) include structured detail:

var valErr *client.HTTPValidationError
if errors.As(err, &valErr) {
    for _, d := range valErr.Detail {
        fmt.Printf("Field %v: %s\n", d.Loc, d.Msg)
    }
}

Streaming

Chat completions and Ollama generate/chat endpoints support NDJSON streaming. The SDK returns two channels: one for events and one for errors.

events, errs := c.StreamChatCompletion(ctx, body)
for {
    select {
    case event, ok := <-events:
        if !ok { return }
        fmt.Printf("Type: %s, Data: %s\n", event.Type, string(event.Data))
    case err, ok := <-errs:
        if !ok { return }
        log.Fatal(err)
    }
}

See examples/streaming/main.go for a complete runnable example with graceful shutdown on SIGINT.

Examples

• examples/basic/main.go — Create a client, call a public endpoint, and handle errors.
• examples/streaming/main.go — Stream chat completions with graceful shutdown.

Buy Me a Coffee

If you find this SDK useful, consider buying me a coffee to support ongoing development.

Thank you!

License

MIT © 2026 notfixingit3