Sider Omni Sidebar - Give every Mac app an Agent Sidebar
byβ’
An AI agent beside any Mac app. Sider Omni sidebar sees what's on your screen and acts on it directly, right where you're working. Watch the result, ask for the next change. Nothing to copy, nothing to upload.
Built on GPT-6 Astra, and it can run on your existing ChatGPT or Codex credits.
Replies
Hi! We're the team behind Sider Omni. π
We built Sider Omni so your agent can join the work you're already doing, with the context to understand it and the ability to act on it.
With a table open in Numbers, ask: βCreate a bar chart of sales by region from this table.β π
Sider Omni reads the table already on screen and builds the chart in Numbers. As you inspect it, keep refining: βNow show return rates by region in a separate chart.β π
You can see the source data and each change in the same workspace, then ask for the next step without exporting data or starting over.
You can also choose from supported models, including GPT-6 Astra, and pick a companion icon that follows you across your apps. β¨
We'd love to hear what you'd use Sider Omni for. π¬
Astra made it possible for us, for the first time, to successfully rewrite Omini Bar's complex Computer Use and Browser Use systems.
These features require coordinating macOS accessibility, visual understanding, browser control, interaction recovery, and tool protocols across a large codebaseβwork that previous models such as 5.6 could not complete reliably.
Astra can correctly execute complex, multi-step development tasks and automate real UI testing, from identifying failures and tracing their causes to implementing fixes and verifying the complete interaction.
This has dramatically reduced development and testing time, enabling us to build capabilities that were previously impractical to complete.
How does it decide which AI model or tool is best suited for a specific task and can users customize those preferences?
Congrats @rick_fan & team!
@rick_fanΒ @hamza_afzal_buttΒ Thanks! You can choose the AI model, while Sider Omni automatically picks the right tools for each taskβsuch as browser actions, desktop app control, file handling, and terminal commands.
@hamza_afzal_buttΒ Thank you, Hamza, you can do a lot of configuration right there in the settings.
Hey everyone π
Omni Sidebar docks to whatever window you're in β Word, Excel, Keynote, Notion, Obsidian, Safari, Mail, PDFs β and follows it when you move, resize, or go fullscreen. No copy-paste, no plugins, no switching apps to ask a question.
Each window gets its own thread, so context stays clean. Pick your own model. And it only reads a window when a task actually needs it β permissions are separate and revocable, and everything runs through a local agent on your Mac.
Mac only for now, and still early. Give it a spin and tell us what breaks β every bit of feedback shapes what ships next.
Would love your support and your upvote today π
Does the data leave your local space?
@busmark_w_nikaΒ The data still needs to be processed by a large language model; you can choose the model you trust. Thank you, Nika
Congratulations on the launch! Having the bot by the side of the application rather than in another separate chat window is the right approach in my opinion.
@hamza_aafzalΒ Thank you! We wanted you to be able to ask for help or get something done without stepping away from what youβre working on.
Really nice. Is it safe to use :)
Also, do i need a local LLM to run this?
@chilaraiΒ Safety is a big priority for us. Sider Omni only reads the screen context needed for a task you start, and the default Ask for approval mode asks for confirmation before important actions. You donβt need a local LLMβyou can use Siderβs model service or connect your own ChatGPT/Codex account.
Love that it acts right on your screen without any copy paste, keeping you in flow instead of pulling you into yet another chat window.
@ilko_kacharovΒ Thank you! We want AI to come to where you work, instead of making you move your work into another chat box. Sider Omni stays beside the current app and helps you get things done right there.
Congrats on the launch , It operates offline or in low band -width conditions?
@easeopsΒ Thank you! AI processing currently requires an internet connection. Text and interface-based tasks are generally lighter on bandwidth, while tasks involving screenshots and visual understanding depend more on the quality of the connection.
I like that the assistant stays close to the context instead of forcing you to jump into a completely separate app.
@salisu_abubakar3Β Thank you! Keeping the assistant beside the current window means fewer interruptions and no need to leave what youβre working on just to get help.