Manus
autonomousCloud AI agent that runs multi-step tasks in a sandboxed environment.
AI agents are software systems that plan and execute multi-step tasks with minimal human input. Unlike single-function AI tools, agents reason over goals, use external tools, and loop until a task is complete. This category lists 485 tools, from business-process automation agents like Workativ to conversational agents built for customer support.
Cloud AI agent that runs multi-step tasks in a sandboxed environment.
Add AI fields and automations to Airtable bases without code
Search and automate workflows across enterprise data
Build AI automations from prompts
AI agents for customer service and collections
Enterprise AI agent for HR workflows
AI agent platform that boosts employee productivity and reduces operational costs
AI phone support that handles customer calls 24/7
Automate sales and support workflows
Automate repetitive work with natural language
Enterprise AI customer support for online stores and service companies
Parallel AI agents across 300+ models for research and automation
Automate tasks on iOS without interrupting your workflow
Manage the full content lifecycle from idea to distribution
AI research agent connected to crypto data sources
Build custom chatbots trained on your own documents
Build and run custom AI agents for any workflow
AI content creation with grammar and plagiarism checks
Manages large language models with monitoring and optimization
Builds tables, charts, and dashboards directly from SQL
AI chatbots for customer service, marketing, and sales
Create AI phone agents without coding or technical skills
AI chatbots and agents for business customer service
Automated blog writing and publishing
The category is broad enough that two tools can both be called agents while working very differently. Some agents are workflow-based, executing a fixed sequence of actions when triggered. Others are goal-directed, meaning they interpret a high-level objective and plan the steps themselves. The latter are more flexible but also less predictable, which matters if you are building a production system. Key things to evaluate: what tools or APIs the agent can access, how it handles errors and ambiguous states, and how much human oversight is built in. Memory and context length also differ significantly between products. For business users, agents that connect to existing SaaS tools (Slack, HubSpot, Jira) tend to deliver value faster than those requiring custom integrations. Pricing structures vary from per-task tokens to monthly platform fees.