Riven Chat
Riven Chat is the conversational workspace at rivenai.io — ask anything and get an answer grounded in live web sources, with citations you can check.
Behind the ask box is a full reasoning workspace: uploads, branching conversations, artifacts, plugins, and fine-grained model control. The service reference is Riven Chat.
You can ask your first question as a guest, no account needed. Signing in unlocks conversation history, uploads, plugins, settings, and higher usage limits — see Account & sign-in.
Cited answers, grounded in the live web#
When a question needs current information, Chat searches the web as part of answering and cites its sources inline. Follow-up questions keep the thread’s context, so you can drill into any claim without re-explaining what you’re working on. The retrieval layer is Riven Grounding.
Choose your model — and how hard it thinks#
The model picker lets you switch models mid-conversation. Beyond picking a model, you can tune its reasoning effort per conversation — lighter for quick lookups, deeper for hard problems. Which models are available depends on your plan tier; the full roster with tier gating is in Models & usage.
Upload files into the conversation#
Attach documents and images directly in the composer. Uploaded files become part of the conversation’s context, so you can ask questions about a report, compare it against something on the web, or carry it through follow-ups.
Branch any conversation#
Any turn can become a fork. Branching lets you explore two directions from the same starting point — a different framing, a different model, or a different reasoning depth — without losing the original thread.
Artifacts: work on documents in-thread#
When a conversation produces something worth keeping — a document, a plan, code — Chat opens it as an artifact beside the thread. Artifacts support inline editing with change tracking, so you iterate on the output itself rather than scrolling back through messages.
Plugins and tools#
Chat supports first-party plugins and open tool integrations, so the assistant can act — not just answer. Tools are configured per-agent in settings; the defaults include search and personal memory. The tool layer is Riven Cortex.
Track your usage#
Your plan balance lives in Settings → Stats: current plan, usage this period, and an upgrade path when you need more headroom. See also Billing & plans.
Chat vs. the API#
| Riven Chat | Riven API | |
|---|---|---|
| Best for | Interactive use — research, drafting, analysis | Building your own product or integration |
| Interface | Hosted web app at rivenai.io | HTTP endpoint, OpenAI-compatible |
| Auth | Your Riven account (guest mode available) | API key (rvn_…) — see Authentication |
| Where | rivenai.io | api.rivenai.io — see API reference |
Next steps#
- Riven Computer — when a question is really a task, hand the whole thing off.
- Getting started — make your first API call.
- Troubleshooting — if something isn’t working.