Rivesa provides an AI-native operating system that turns recurring work into persistent agent loops, accessible through a conversational interface. The platform gives each AI agent a "passport" with its own identity, permissions, credentials, and activity ledger, enabling trustworthy delegation of real work across shared channels and messaging surfaces like iMessage.
Funding
Funding not disclosed
Founders
Product
Problem
Teams face a growing mismatch between workload and headcount, with recurring operational tasks consuming time that could go to higher-value work. Existing AI tools require users to learn new interfaces, often operate with a single set of credentials that limits collaboration, and lack the accountability mechanisms needed to trust agents with real responsibilities.
Solution
Rivesa provides an AI-native operating system where a single assistant interface sits in front of a team of specialized agents that execute workflows across channels. Each agent carries a "passport"—a persistent identity with defined roles, skills, authorized accounts, and a full ledger of its actions and costs—so delegation is governed by enforced boundaries rather than prompt instructions. The platform integrates with messaging surfaces like iMessage, allowing users to text the assistant directly and receive human-like responses, while per-person OAuth connections ensure each participant's data is accessed according to their own permissions. Managers can monitor agent progress, review activity maps, and coordinate multi-step workflows through shared channels.
Target Audience
Primary users are individuals and small-to-medium teams looking to automate recurring operational workflows, from personal assistants to departmental coordinators managing outreach, scheduling, and data management tasks.
Features
- Passport-based agent identity system with name, role, credentials, skills, authorized accounts, and memory
- Immutable ledger recording every agent run, including actions taken, duration, and cost
- Enforcement of authorizations at the system level, making agent boundaries a property of the agent rather than a prompt suggestion
- iMessage integration for conversational interaction with no separate app or signup required
- Per-person OAuth credential flow, granting each participant access to their own calendars, email, and files without exposing passwords
- Shared channels where multiple agents and humans collaborate, with visible tool progress and workflow coordination
- Read-only capabilities for information lookups, with explicit confirmation required for any action that changes data
- Connector setup for email, calendar, and other external accounts, with the fallback that unanswerable messages wait and are delivered once tools are available