Understand the project
Practical overview
When it fits
Use Agentao when an agent must be hosted in Python and its permissions, inputs, memory, protocols or sessions need explicit runtime controls. Its CLI path suits terminal operation, while embedding and ACP support address application and agent-service integrations.
Who it is for
- Python developers embedding AI agents into host applications.
- Developers running agents from a terminal.
- Builders exposing full agents through ACP.
- Teams needing explicit permission modes, persistent memory or protocol-based tools.
Documented uses
- Embed an agent runtime into a Python application.
- Run an agent through a CLI.
- Expose an agent through an ACP server.
- Control tool access with permission modes and confirmation.
- Add persistent user or project memory.
Before you choose
Requirements & limits
The repository is public and MIT-licensed. The documented embedding installation uses agentao, while the CLI uses the cli extra. The CLI quickstart requires an OpenAI-compatible API key, base URL and model.
- Python is the primary implementation language and the documented embedding path targets Python hosts.
- The runtime supports four permission modes: read-only, workspace-write, full-access and plan.
- MCP supports stdio, Streamable HTTP and SSE according to the feature overview.
- Persistent memory uses SQLite and separates user and project scopes.
- The repository documents CLI, embedded and ACP server interfaces, but does not provide deployment configuration details in the supplied material.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Evidence and transparency
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 3 October 2026
- Original website title
- jin-bo/agentao · Official homepage title · retained in its source language