A self-improving AI agent that creates and refines skills from experience, runs on many model providers, and works through CLI and messaging apps.
shiwenwen/hope-agent
A local-first desktop AI assistant with long-term memory, autonomous goals, dynamic workflows, and a headless server mode for NAS or cloud hosts.
Open-source alternative to:ChatGPT
- Stars
- 1,852
- Forks
- 168
- Open issues
- 66
- Last commit
- 2 days ago
What is it?
Hope Agent is a desktop AI agent built with Rust and Tauri that keeps sessions, memory, and task state across a desktop app, a browser-based Web GUI, and chat apps such as Telegram, Discord, Slack, and Feishu. It can also run as a headless server (hope-agent server) on a NAS, home server, or cloud VM, with the full Web GUI embedded in the binary.
The project covers more than chat. Goals hold an outcome and completion criteria, Workflows organize multi-step execution with traces, and Loops decide when to resume work on a timer or event. Memory is split across global, project, and agent scopes, with on-demand recall and skills drafted from complex work. It can also control the computer and browser on macOS, call MCP tools, and run commands or file edits, with tool approvals and an optional Docker sandbox guarding execution.
The desktop app is macOS-complete, with Linux and Windows marked experimental. Model access covers about 50 provider templates, plus one-click local models via Ollama. Data is stored under ~/.hope-agent/ by default, and API calls go directly to model providers.
Key features
- 01Cross-device session handoff between desktop, browser, and IM channels
- 02Goal, Workflow, and Loop for long-running, resumable tasks
- 03Cross-session memory organized by global, project, and agent scope
- 04Tool approvals with optional Docker sandboxing for risky commands
- 05Headless server mode with an embedded Web GUI for NAS or cloud hosts
- 06MCP client support and lifecycle hooks
Use cases
- Running a always-on personal assistant on a home NAS and reaching it from a phone browser
- Continuing a research or drafting task started on desktop from a Telegram or Feishu chat
- Managing a project with shared memory, uploaded files, and scheduled recurring work
- Using the agent as an ACP backend for an IDE
Topics
FAQ
- What is hope-agent?
- A local-first desktop AI assistant with long-term memory, autonomous goals, dynamic workflows, and a headless server mode for NAS or cloud hosts. Hope Agent is a desktop AI agent built with Rust and Tauri that keeps sessions, memory, and task state across a desktop app, a browser-based Web GUI, and chat apps such as Telegram, Discord, Slack, and Feishu. It can also run as a headless server (hope-agent server) on a NAS, home server, or cloud VM, with the full Web GUI embedded in the binary.
- Is hope-agent free to use? What license does it have?
- hope-agent is open source under the MIT license. Check the license terms before using it in commercial products.
- Is hope-agent still maintained?
- Actively maintained. The last commit to hope-agent was on Oct 20, 2026.
- What can I use hope-agent for?
- Running a always-on personal assistant on a home NAS and reaching it from a phone browser Continuing a research or drafting task started on desktop from a Telegram or Feishu chat Managing a project with shared memory, uploaded files, and scheduled recurring work Using the agent as an ACP backend for an IDE
- Where can I get hope-agent?
- The source code and installation instructions are on GitHub: https://github.com/shiwenwen/hope-agent
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Open-source personal AI assistant and Agent Harness that plans tasks, runs tools and skills, and keeps memory across web and IM channels.
A self-hosted personal AI assistant in Rust with WASM-sandboxed tools, encrypted credential storage, and a web UI, inspired by OpenClaw.
A self-hosted personal AI agent server in a single Rust binary, with sandboxed tool execution, multiple LLM providers, and chat channels like Telegram and Discord.
Self-hostable personal AI assistant as a single Go binary, combining LLM chat, RAG knowledge bases, and agent loops for browsing, shell, and MCP tools.
Self-hosted personal AI agent framework in Python with a browser WebUI, terminal client, chat app channels, tools, memory and MCP support.