The web app at vm0.ai is the default home for Okou. It's where most teams do their heaviest work — research projects, content drafts, multi-step work, anything that benefits from a wide screen and a full chat log on the side.
If you're going to do one thing in Okou, do it in web chat. Once you're comfortable, the other channels make sense as specialized tools.
What web chat is best for
The kinds of work that really shine here:
- Research with citations. A 20-tab browsing task condensed into one prompt: "Brief me on Acme Corp — funding, news, prior interactions, open opportunities." Okou pulls from CRM, email, news, and produces a brief with inline links.
- Long drafts. Blog posts, PRDs, board memos. Okou drafts, you edit in the same chat, Okou revises.
- File-heavy work. Drop a PDF, a CSV, or a screenshot and ask Okou to work with it. "This is last quarter's expense report. Find anomalies and group them by department."
- Multi-chat days. Open a research task in one tab, a content draft in another, a debugging chat in a third. Each is isolated; switching is
Cmd+K. - Anything you want to iterate on visually. Tables, charts, code blocks, and diagrams render inline. You see the artifact in context.
The interface, at a glance
- Left sidebar — your chats, grouped by recency, with search. New chat at the top.
- Center — the active chat. Your prompt, Okou's response, and the run log (toggleable).
- Side panel (when active) — one panel shared by artifacts, email drafts, browser sessions, and automations, so whatever you're inspecting sits beside the conversation instead of replacing it.
- Sidebar → Artifacts — every deliverable Okou has produced for you, grouped by type. See Artifacts.
- Top-right — workspace switcher, settings, billing.
The whole app is keyboard-driven. The shortcuts worth learning early:
Cmd+K— jump to any chat or workflowCmd+N— new chatCmd+/— show all shortcutsShift+Enter— newline in the prompt;Entersends
File uploads
Drag a file into the chat input and Okou ingests it before responding. Supported formats:
- PDFs (text and scanned — Okou OCRs)
- Images (PNG, JPG, WebP, HEIC)
- Audio (MP3, WAV, M4A — transcribed automatically)
- Video (MP4 — transcribed; key frames available to the model)
- Office documents (DOCX, XLSX, PPTX)
- CSV, JSON, Markdown
- Code archives (ZIP, tar.gz)
Large files take a few seconds to ingest; you'll see a progress indicator. The file is available to Okou for the rest of the chat.
Multi-chat work
Open multiple chats in tabs (or pop them out into windows). Each has its own context — Okou in tab A doesn't see tab B unless you hand it over. This is the right mental model for:
- Doing research in one tab while drafting in another
- Running an expensive long task (a market study) in the background while continuing other work
- A/B testing two prompts against the same task
The sidebar's recent-chats list shows everything you've opened, with status indicators if a chat is still working.
Bringing another thread in
When the next task depends on work you did somewhere else, you don't have to re-explain it. Type @ in the composer to pull in a previous chat thread — the menu lists titled threads from across your agents, each with its agent's avatar. Pick one and it goes into your prompt as context, ready for "compare this with…", "continue where that left off," or "review what we decided there."
The current agent reads what you handed it. It doesn't inherit the other agent's connectors or permissions — see Agents.
Steering a run
You don't have to wait for a run to finish before correcting it. Send the correction while Okou is working and it's picked up inside the active run, rather than throwing away what's already done. To reply to one exact phrase, select it and choose Quote to drop that text into your next message.
Diagrams
Ask for a flowchart, a sequence, a process map, or an architecture view and the result renders as a diagram in the conversation rather than as a block of diagram code. Open it beside the chat when you want more room, then come back without losing your place.
Voice input
You can dictate instead of typing. Voice input is metered per member and the allowance resets daily — Free includes a small lifetime trial, and Pro and Team raise both the request count and the minutes per day. See Credits & billing for the current limits.
Working in your language
Pick a language under Settings → Preference and the whole product follows: chat, agents, workflows, connectors, artifacts, billing, onboarding, down to the error messages. Ten languages are available — English, German, Spanish, French, Italian, Japanese, Korean, Hindi, Indonesian, and Brazilian Portuguese.
Okou starts from your browser's language on first visit, and your choice is saved to your account rather than to the machine, so it follows you to another computer.
Chat history and search
Every chat you've ever run is in the sidebar, searchable. Two patterns:
- Find a past artifact. "Where's that brief I had Okou write about Acme in March?" Search the sidebar; the chat shows the brief inline.
- Resume a finished chat. Click any past chat and continue the conversation. Okou re-loads the full context.
Search is full-text over your prompts, Okou's responses, and the artifacts produced.
When the work is on a web page
Research and scraping are handled by managed web services with nothing to connect. When a page needs a real browser — a form to fill in, a portal you're signed in to, a site with no API — enable Cloud Browser for that chat and watch, guide, or take over from the same side panel.
When to use a different channel
- For team-visible work in a channel context → Slack.
- For a personal or group bot outside Slack → Telegram.
- For quick replies and approvals from your phone with no app → Phone.