
AgentScope is an open source Python framework for developers building AI agent applications on their own servers. It combines an agent SDK with a FastAPI backend and web UI, including support for separate users and sessions and distributed deployment. Its Apache 2.0 license allows use and modification.
Agents use ReAct reasoning to choose tools and act on their results. They can return structured output, pause and resume during execution, and run tools sequentially or concurrently according to the tools' properties. Built-in coding tools cover shell commands, file editing and search; agents can also use Python functions, MCP servers and skills.
Tool access has several controls: review for individual tools, human approval and isolated execution. Sandboxes can use a local filesystem, Docker, E2B or Kubernetes, with isolation at the user, agent or session level. Middleware lets developers adjust agent and tool behavior without rewriting the agent itself.
AgentScope supports Ollama for local LLM use, alongside cloud providers including OpenAI, Anthropic, Gemini and DeepSeek. The application backend can be self-hosted; model requests go to the selected local runtime or cloud provider.
For longer tasks, context management compresses conversations, moves bulky tool results out of context and retrieves relevant information. It supports RAG and long-term memory, with Mem0, ReMe and vector database integrations. The web UI receives streamed reasoning, tool calls and text, image and audio content.
Claim this page with an email at docs.agentscope.io. AgentScope gets the verified badge, and you can upgrade the listing to be featured on localhosted. Proud to be listed? Put our badge on your site.
Want more people to find AgentScope?Promote it
Something wrong or outdated on this page?
2.6KUpdated 2 days agoApache-2.0
#Distributed execution#Guardrails#Human approval
AgentField is a self-hosted, open source platform for developers building AI agents that need to work together under shared access rules. It runs in your own infrastructure and combines agent execution with identity, policy enforcement and audit records. The project uses the Apache 2.0 license.
447Updated 4 weeks agoApache-2.0
#Agent Skills#Code execution#Guardrails
42.5KUpdated 12 hours agoApache-2.0
Docker · Web#Guardrails#Human approval#LLM tracing
1.5KUpdated 2 days agoMIT
Docker · Web#Code execution#Context compression#LLM tracing
711Updated 1 month agoMIT
Web#Agent Skills#Code execution#Git integration
28.5KUpdated 7 hours ago
Web#Human approval#LLM tracing#MCP
Hypha is a local-first TypeScript framework for developers building AI agents that need controlled tool use, recoverable workflows and an audit trail. It separates the agent's reasoning and planning from the execution layer that enforces policies, approvals and workflow boundaries. It's open source under Apache 2.0.
Agno is a Python framework and runtime for developers building customer-facing or internal AI agents. You can run its agent platform locally with Docker, on your own servers or in your cloud. The open-source framework uses the Apache 2.0 license, and the platform keeps sessions, memory, knowledge and traces in your database.
Exo Agent Framework is a self-hosted framework for developers building long-running AI agents that can revise their own code, tools and prompts. Its exoharness foundation keeps agent history and resources separate from the logic that decides what the agent does next. The project is open source under the MIT license.
GitAgent is an MIT-licensed, open-source AI agent framework for developers who want to keep an agent's behavior and memory under version control. It runs locally, with its configuration, personality, rules and saved memory stored as files in a Git repository. That makes changes reviewable through diffs and history, and lets teams fork an agent or maintain different behaviors on separate branches.
Mastra is a TypeScript framework for developers building AI agents and applications on their own servers or inside existing web apps. Its server runs locally or as a standalone deployment; Mastra Cloud provides a hosted alternative. Model routing connects to providers such as OpenAI, Anthropic, and Gemini, so running the framework locally doesn't keep those model requests on your machine.