Oasis
Mercury provides a persistent collaborative environment where humans and AI agents work together in shared rooms on live documents, spreadsheets, and codebases. Agents retain knowledge across sessions, eliminating cold starts, and can be shared across teams as reusable, specialized teammates. The platform integrates with Slack, iMessage, Teams, and WhatsApp for seamless team communication.
- AI Agents
- Artificial Intelligence
- Enterprise Software
- Developer Tools
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
Teams working with AI agents often face fragmented workflows where agents operate in isolation, lack shared context, and require time-consuming setup for each new session or model. This leads to cold starts, inconsistent outputs, and difficulty scaling agent usage across an organization.
Solution
Mercury provides a persistent collaborative environment where humans and AI agents work together in shared rooms on the same live documents, spreadsheets, and codebases. Agents built within Mercury retain knowledge from every run, allowing teams to swap or upgrade agents without losing context or momentum. The platform supports shared agent fleets, where specialized agents can be created once and reused across the team, connecting to the services and apps members use daily. Automatic routing and built-in human oversight keep workflows efficient while ensuring agents stay aligned with team goals and quality standards. Agents have their own computers and can work autonomously through tasks like onboarding conversion plans, while joining team conversations on Slack, iMessage, Teams, and WhatsApp.
Target Audience
Product and engineering teams at forward-looking companies who want to integrate AI agents directly into their daily collaborative workflows and build reusable agent fleets, including teams in legal, sales, healthcare, and trading compliance operations.
Features
- Room-based group chat where multiple humans and agents collaborate simultaneously in one workspace
- Live co-editing of documents, spreadsheets, and codebases with both human and agent participants
- Persistent agent memory that carries context across model swaps or agent replacements, eliminating cold starts
- Shared agent library enabling team-wide reuse of specialized, pre-configured agents
- Automatic task routing between human and agent team members based on workflow context
- Human-in-the-loop controls to monitor and approve agent actions in real time
- Agents with their own computers that work autonomously through multi-step plans, marking progress as they complete tasks
- Integration with Slack, iMessage, Teams, and WhatsApp for joining team conversations