Open WebUI

The freedom AI stack.

[Open WebUI @OpenWebUI]

One interface. Every model. That's the idea behind Open WebUI + OpenRouter.

Instead of juggling multiple vendors, API keys, and UIs, organizations can now get a single platform with 600+ models ready to go out of the box.

Interface and inference, together.

[0xBakeer @0xBakeer]

First impression of how fast the new Qwen3.8-Flash-Next runs on a single DGX Spark.

180B params, 262k context, all local.

[Christopher Man @int32max]

setup and self hosted codex for my entire family (8 family members)
- @OpenWebUI running harness and environment (docker container)
- @concentrateai router providing model (deepseek v4 flash)
- @Cloudflare dns + tunnel + access, with https handled at Cloudflare’s edge
- @ChatGPT + @Tailscale remote session and make changes to the docker container
- old gaming desktop running debian in my flat in london

this cost me basically nothing but some electricity and credits on @concentrateai which I am migration from @OpenRouter today

making it easy for my family to become ai pilled in a way where I can help guide them - basically setup enterprise ai for my family

helps out a ton by the fact I can setup skills and mcp connectors for my family and they dont have to worry about what skills are good etc

will make a youtube video soon about how @Cloudflare made this setup super fucking easy with cf access helping out with email otp and tunneling

[Ash Hart @ashxhart]

iPad connected over WiFi AP to the Spark running DeepSeek V4 Flash.

Imagine rolling up to the coffee shop with this naughty little setup.

Harness on the iPad time 👀😅

@NVIDIAAI @vllm_project @OpenWebUI @Apple

[0xBakeer @0xBakeer]

Local AI is very practical.

One prompt: download this arXiv paper PDF and save it into my AI Knowledge. Then summarize the core idea.

That is a real workflow. A tool call, a fetch, a file written into my knowledge base, then a read and a summary. Not a toy prompt.

Running deepseek-v4-flash locally in Open WebUI, with tools and web access on. Start to finish, from my first message to the finished summary, two minutes. I recorded it unedited so nobody has to take my word on the speed.

That is the part people still get wrong about running models on your own hardware. The assumption is that you trade away everything for privacy. Slow tokens, weak reasoning, no tools. That was true a year ago. It is not true now.

No API bill. No rate limits. The paper never touched a third party server, and neither did anything I asked about it.

Two minutes, on my own box (dgx spark).

[OpenRouter @OpenRouter]

Interface + inference, in one place.

@OpenWebUI now runs on OpenRouter.

Give your team one chat interface, one unified bill, and access to 400+ frontier and open models through a single API.

[Open WebUI @OpenWebUI]

https://t.co/sWotsijyPL

[Open WebUI @OpenWebUI]

One interface. Every model. That's the idea behind Open WebUI + OpenRouter.

Instead of juggling multiple vendors, API keys, and UIs, organizations can now get a single platform with 600+ models ready to go out of the box.

Interface and inference, together.

[いー @EXOR2327209]

open webui初め���触ったけどすごいなこれ
コンテンツ量おおすぎる

[Docker @Docker]

Generating images shouldn’t mean API keys, credits, or sending prompts to the cloud.

This shows how to run image generation locally with Docker Model Runner + Open WebUI, with a chat UI and API on your own machine.

No cloud required.

Read → https://t.co/RWoKN006U5

[Tom Dörr @tom_doerr]

Open WebUI native app without Docker setup

https://t.co/f6h8G7xopb

[Ivan Fioravanti ᯅ @ivanfioravanti]

Another super trick for Hermes Agent, you can create multiple profiles/gateway.

Connect each of them to Open WebUI and grant access to your family members to their own Hermes Agent.

Thanks me later.

[透木 明@-(TRYPTOPHAN)- @nemutteiruneko]

OpenwebUIにヤバい機能が追加された！！！
・カレンダー
・オートメーション

[Omarchian Moehre @Moehrenmann]

Seit knapp 1 Jahr meine Lieblingsoberfläche für KI. Und sie wird besser und besser. Nun wurden Features hinzugefügt, die Openclaw als Erstes aufwies und die ich mir seitdem in Open-WebUI gewünscht habe.
https://t.co/ln7kymt4z1

[Geek @geekbb]

Open WebUI 大版本更新到 V9.0

• 上了官方桌面版，Mac / Windows / Linux 都能直接用
• 加了 AI 定时自动化，日报、提醒、周期任务都能跑
• 新增日历和任务管理，不只是聊天工具了，开始往 AI 工作台方向走了
• 强化了 Azure OpenAI、Ollama、Responses API 支持，工具调用更完整
• 做了大量性能优化，长对话、文件处理、流式输出都更顺。

[ClassicMain @ClassicMain]

.@OpenWebUI - Open WebUI 0.9.0 just dropped.

Desktop App. Scheduled AI Automations. Calendar. Task Management. FULL ASYNC BACKEND REWRITE.

200+ changes. The biggest release EVER.

Everything is faster. Everything.

https://t.co/fVBKpr6oXw

[XDA @xdadevelopers]

After two months of Open WebUI updates, I'd pick it over ChatGPT's interface for local LLMs
https://t.co/1L8BN6zH3v

[Open WebUI @OpenWebUI]

0.9.0, take it or leave it.

https://t.co/Z2pwqn4daW

[ClassicMain @ClassicMain]

Shipped something that shouldn't work:

An @OpenWebUI plugin that makes your local model stream interactive SVGs, dashboards, and charts directly into chat — painted live, token by token, as it generates

no core patches. one tool + one SKILL.md.

https://t.co/swn4N6491n

[Dave Lee @davelee]

This afternoon I picked up a new Nvidia DGX Spark computer with the goal of trying to run Gemma 4 31b (4bit) on it locally as a server.

Just 1.5 hours later, it’s working!

Using Open WebUI on my MacBook as the interface and it’s connecting to my DGX Spark running as a Gemma 4 server.

[Teknium 🪽 @Teknium]

Hermes Agent tip of the night - did you know with the OpenAI Endpoint that Hermes Agent create for itself, you can use @OpenWebUI to have a chat GUI for your agent?

See the full guide here:

https://t.co/CYpcPo9n2F