Durable context per DM, group, workflow lane, or specialist agent.
/ 01 - Intro
Operate Your Agents.
Ravi connects WhatsApp, Telegram, Discord, and Slack through omni, then runs work through Codex, local coding agents, and more runtimes soon.
Install / macOS + Linux
bun install -g ravi.bot@latest
/ 02 - Surface
Work Routing Preview
The public shape is simple: messages enter, Ravi resolves context, then the right runtime surface takes over.
Incoming Intent
turn this conversation into a tracked task
Permissions, scope, executable checks, and runtime boundaries stay explicit.
When work becomes execution, it gets a task, artifacts, status, and follow-through.
/ 03 - Workflow
Four Operating Layers
01
Capture
WhatsApp, CLI, overlay, and background events enter the same operational stream.
02
Route
Ravi resolves session, agent, permissions, model, and execution context before work starts.
03
Execute
Specialized agents work inside scoped runtime lanes with task artifacts and clear state transitions.
04
Report
Results, denials, insights, tasks, and workflow status return to the operator surface instead of disappearing into logs.
/ 04 - Runtime
Stable Core.
Open Surfaces.
Ravi is the harness. The visible UI can change; routing, tasks, permissions, and coordination stay coherent.
sessions -> agents -> tasks -> workflows
Codex / local coding agents / more soon
WhatsApp / Telegram / Discord / Slack
whatsapp overlay / cli / cockpit / future apps
permissions + audit + hooks + model routing
workflow coordinates -> task executes -> session reports
/ 05 - Install
Start From The CLI
bun install -g ravi.bot@latest
ravi setup
ravi doctor
ravi daemon start
ravi daemon status
ravi agents list
ravi agents show main
ravi sessions list
/ 06 - Close
The work stays visible.
Ravi keeps context durable, routes the next action, protects boundaries, and brings the work back to the operator surface.