General-purpose AI agents
Manus
Manus builds general-purpose AI agents that research, analyze data, write code and deploy applications for individuals and business teams. Users can build full-stack web apps with databases and payments, or produce research, slides and charts. Its Team plan adds shared projects, integrations with tools such as Google Calendar, GitHub, Notion and Slack, and administrative controls. Per-seat subscriptions pool usage credits across a team, with additional credits available to purchase.
Launched in 2025, Manus counts Red and Pan among its founders; Red served as CEO in March 2026. Its engineering approach builds on frontier models through context engineering. Agents execute tools in virtual-machine sandboxes and feed observations into subsequent decisions. Its published engineering lessons emphasize preserving prompt-cache reuse and using the file system as persistent memory, allowing information to leave the immediate model context while remaining retrievable.
In December 2025, the company reported more than $100 million in ARR and a total annualized revenue run rate above $125 million, including usage-based and other revenue. Meta acquired Manus on December 29, 2025. On August 11, 2026, Manus announced that it would return to independent operations as part of its separation from Meta; that announcement described a transition in progress, not a completed separation.
1 talk
Newest first1 speaker at AIE
Affiliations reflect their AIE appearances, not necessarily current employment.
Messages from the stage
Task processing and conversation context
The workshop covers polling asynchronous tasks, uploading files, and processing webhooks with Modal. A Slack bot example shows how task context can be maintained across conversation threads.
Affiliations reflect each recorded session, not necessarily current employment.
