DEV Community

Cover image for UNO: I gave one chat a whole team of AI specialists. It hands back real files, not text.
Jagdish Chudasama
Jagdish Chudasama

Posted on

UNO: I gave one chat a whole team of AI specialists. It hands back real files, not text.

Try it now → uno.nodeops.network

Most AI tools quietly push the hard part onto you. You pick the model. You pick the tool. You copy the output into the next tool. You stitch the results together yourself. The orchestration is your job.

UNO removes that step.

It is an AI workspace. A team of specialists. One chat.

You ask in plain language. An orchestrator routes the request to the best-fit specialist across engineering, design, data, finance, legal, product, writing, and marketing. Bigger asks pull in several specialists who collaborate. You still get one consolidated reply in one thread.

What makes it useful in practice, not just in a demo

A request fanning out to several specialists, then one reply

It returns real files. It writes and runs code, then hands back finished PDFs, Word docs, spreadsheets and slide decks. Not a wall of text you have to rebuild. Ask for "a go-to-market plan as a PDF" and the PDF shows up inline and in the Files panel.

Finished files in the Files panel — PDF, DOCX, XLSX, PPTX

It acts inside your accounts, behind a gate. Connect Gmail, Calendar, Notion. Before it does anything external, it shows you exactly what it is about to do and waits for your approval. Every action lands in an action history log. No silent sends.

Approval gate before an external action

You control cost and models. Open the Team page and pick which model each specialist runs. A faster, cheaper model for simple work. A more capable one for the hard work. Swap it mid-conversation, no redeploy, effective on your next message. Every reply shows the per-turn and per-model spend.

It remembers. Within a chat, and across sessions, scoped to your account.

It reaches you where you already chat, and runs on a schedule. Connect Telegram, Slack, or Discord from the Channels page and talk to your team there. Ask it to run a prompt on a schedule and deliver the result on its own.

Channels — Telegram, Slack, Discord

It is honest when something fails

UNO verifies the files it produced and reports back if something did not work. Document generation can verify-and-re-run when the first attempt does not produce the file. You also get web search with cited sources built in, you can upload your own files for it to read, and after a restart past sessions and files restore from cloud backup.

Honest about scope

So you are not surprised:

  • Text and files only. You can dictate input by voice, but UNO replies in text and files - no spoken replies, no audio generation.
  • No image generation on the built-in models yet.
  • No video generation.
  • No computer or browser control.
  • No native mobile app. UNO is web, plus Telegram, Slack, and Discord.

One more, for the cost-watchers: the orchestrator's own model cost is not itemized separately yet. Specialist costs are shown per model.

Try it

Deploy your UNO agent, open Chat, and ask it something in your own words. The setup is deploy and go. You do not wire up tools or pick specialists just to get started.

→ Start at uno.nodeops.network

Build, run, scale, and monetize. One place.

Top comments (0)