Coding & Development✨gemini🤖claude💬gpt🌐universal
Vizier
Right-hand Agent for your majesty;
Quick Start
Install this Skill
npx skills add vizier-lab/vizier# Vizier > 21st Century Digital Steward; Right-hand agent for you majesty Vizier is a Rust-based AI agent framework that provides a unified interface for AI assistants across multiple communication channels (Discord, Telegram, HTTP, WebUI) with memory, tool usage, and extensible architecture. ## Features - **Multi-Channel Support**: Connect to Discord, Telegram, HTTP (REST API & WebSocket), and WebUI - **AI Model Integration**: Support for multiple AI providers (DeepSeek, OpenRouter, Ollama, Anthropic, OpenAI, Gemini, Xiaomi MiMo, Llama.cpp) - **Memory System**: Session-based short-term memory, configurable recall depth, and vector-based long-term memory - **Tool System**: Extensible tool framework including shell execution, web search (Brave Search), HTTP client, web fetch, scheduler (cron & one-time tasks), vector memory, workspace document management, sub-agent spawning, Python sandbox, and inter-agent communication - **Scheduler**: Built-in task scheduler for automated agent execution - **WebUI**: Modern React-based web interface for interaction and management - **Configuration Driven**: YAML seed config with runtime management via WebUI ## Installation and Configuration ### Prerequisites No prerequisites required for standard installation. The install script handles everything automatically. #### For Custom Installation (Building from Source) - [Rust and Cargo](https://rust-lang.org/) installed ### Quick Start 1. **Install Vizier** (Recommended): ```sh curl -fsSL https://get.vizier.rs | sh ``` Or install via cargo (requires Rust): ```sh cargo install vizier # Or using cargo-binstall (faster) cargo binstall vizier ``` 2. **Generate configuration and workspace:** ```sh vizier onboard ``` This will walk you through provider selection, embedding config, storage backend, and HTTP server setup. 3. **Run the agent:** ```sh vizier run ``` 4. **Open the WebUI** at `http://localhost:9999` to create and manage agents. ### Development Setup For development, clone the repository and use the provided `just` commands: ```sh # Install dependencies (Rust crates and webui npm packages) just install # Run in development mode with hot-reload just dev # Build the webui just build ``` See the [Justfile](Justfile) for all available commands. ### WebUI The web interface is built with React and served automatically when the HTTP channel is enabled. After building (`just build`), it will be available at `http://localhost:9999` (or the port configured in your `config.yaml`). ## Update Installed Version ### Using Install Script Simply re-run the install script to get the latest version: ```sh curl -fFSL https://get.vizier.rs | sh ``` ### Using Cargo (if installed via cargo) 1. Install `cargo-update` if you haven't already: ```sh cargo install cargo-update ``` 2. Update the binary: ```sh cargo install-update vizier ``` ## Planned Features (V1.0.0) - [x] Web UI (React-based interface) - [x] Scheduler and task system (cron & one-time tasks) - [x] Vector memory for long-term retention - [x] Brave Search integration - [x] Local embedding model support - [x] Docker Sandbox - [x] Additional AI providers (Google Gemini, OpenAI, Anthropic, Xiaomi MiMo, etc.) - [x] Sub-agent spawning for parallel task execution - [x] Model Context Protocol (MCP) integration - [x] Skill system for reusable agent behaviors - [x] Built-in HTTP client tool - [ ] WASM-based plugin system ## Development ### Project Structure - `src/`: Rust source code - `agents/`: Agent process loop, LLM interaction, tools, hooks, skills - `channels/`: Discord, Telegram, HTTP (REST + WebSocket + WebUI serving) - `storage/`: Filesystem and SurrealDB storage backends - `config/`: YAML seed config deserialization - `schema/`: Shared types (responses, agent IDs, provider entries) - `mcp/`: MCP client + server integration - `embedding/`: Local embedding models (fastembed) - `shell/`: Shell execution abstraction (local + Docker) - `scheduler/`: Cron and one-time task scheduler - `transport/`: Command transport for agent/channel/global commands - `webui/`: React-based web interface (React Router v7 + Tailwind v4 + TypeScript) - `templates/`: Template files for agent configuration and identity - `.vizier/`: Workspace directory for runtime data (config, database, agent workspaces) ### Available Commands See the [`Justfile`](Justfile) for available commands: | Command | Description | |---------|-------------| | `just install` | Install all dependencies (Rust crates + webui npm packages) | | `just dev` | Run in development mode with hot-reload | | `just run` | Run in attached mode | | `just release` | Build release binary | | `just docker` | Start Docker services (database, etc.) | | `just build` | Build the webui frontend | ### CLI Commands The `vizier` binary provides these subcommands: - `vizier run [--config <path>]`: Start agents, server, and channels (runs ... (truncated)
Repository
10stars
2forks
Details
Related Skills
More coding & development skills you might like
Reticle
Coding & DevelopmentAI agents can generate code, but still struggle to understand what they build. Reticle brings Jev-style machine-native runtime perception to web & desktop applications.
🤖claude🌐universal
1.2k
190
View
agentscope-java
Coding & DevelopmentExpert Java developer skill for AgentScope Java framework - a reactive, message-driven multi-agent system built on Project Reactor. Use when working with reactive programming, LLM integration, agent orchestration, multi-agent systems, or when the user mentions AgentScope, ReActAgent, Mono/Flux, Project Reactor, or Java agent development. Specializes in non-blocking code, tool integration, hooks, pipelines, and production-ready agent applications.
🌐universal
6.0k
1.5k
View
ospec
Coding & DevelopmentDocument-driven OSpec workflow for initialization, change/goal routing, validation, archiving, and durable project knowledge.
✨gemini🤖claude💬gpt🌐universal
524
34
View
misakanet-failure-memory
Coding & Development📚 A zero-dependency, git-backed micro-lesson library for AI Agents to asynchronously share and search verified debugging experience. Python stdlib only. | https://misakanet.org
🤖claude⌨️cursor🌐universal
424
158
View