Perplexity Introduces Teammates for Multi-Agent Collaboration
AI assistants · AI productivity
Perplexity has announced Teammates, enabling users to create multiple persistent AI agents for long-term projects and inter-agent task delegation. Built on its Computer platform, the agents feature long-term memory for project history and can integrate into Slack channels. The feature is currently available to Perplexity Max web subscribers.
Perplexity announced the release of Teammates on November 4, allowing users to deploy persistent AI agents capable of managing long-term projects. These agents can proactively track progress, execute tasks without continuous prompting, and hand off tasks among themselves. Built upon Perplexity's existing Computer multi-model agent platform, Teammates agents can access relevant tools, shared files, and retain memories of goals and past feedback. Teams can also invite agents into Slack channels for task coordination. Operationally, each agent runs in an isolated environment with separated browser sessions, allowing users to inspect or intervene at any stage. Access is currently limited to Perplexity Max subscribers on the web.
Key takeaways
- Perplexity launched Teammates to enable persistent AI agents to handle specialized tasks, track progress proactively, and coordinate workflows.
- Agents retain project memory and context, share files, and can be integrated into Slack channels for team interaction.
- Each agent operates in an isolated environment, with initial web access rolled out to Perplexity Max subscribers.
Source：iThome Read the original ↗
Prepared, checked and automatically published by AI; not individually reviewed by a human. Translation checked by AI.