# Mysti - Your AI Coding Team Working Together

Your AI Coding team for VSCode
12 AI providers — Claude Code, Codex, Gemini, Copilot, Cline, Cursor, OpenClaw, Manus, OpenCode, Qwen Code, Ollama & LocalAI — working solo or in teams
Wisdom of the crowd where the collective intelligence of several agents outperforms a single one.

Providers • Brainstorm • Features • Quick Start • Config • Docs

--- ## What's New in v0.4.0 ### 4 New Providers (12 Total) Mysti now supports **12 AI providers** — added **OpenCode**, **Qwen Code**, **Ollama**, and **LocalAI** alongside Claude Code, Codex, Gemini, GitHub Copilot, Cline, Cursor, OpenClaw, and Manus. Run local models with Ollama/LocalAI or use cloud providers like OpenCode and Qwen Code. Each provider has its own authentic logo in the UI. ### Qwen Code Alibaba's AI coding CLI with deep reasoning capabilities. Uses the same streaming protocol as Claude Code for seamless integration. Supports Qwen3 Coder models with plan, auto-edit, and yolo approval modes. ### OpenCode Multi-backend coding agent supporting Anthropic, OpenAI, Google, and Groq through a single CLI. Uses your configured default model — no lock-in to specific providers. ### Local AI Support Run AI models locally with **Ollama** and **LocalAI** — no cloud subscription needed. Full privacy, zero latency, complete control over your models. ### Stability & Testing 18 stability fixes across brainstorm mode and @-mention tagging — silence-based timeouts, auth pre-checks, convergence guards, mention limits, and retry cleanup. Plus **360 automated tests** via vitest ensuring reliability across all features. ### Bug Fixes - **Windows**: Fixed `spawn EINVAL` error with auto shell mode (#14) - **Codex**: Brainstorm now respects `mysti.codexPath` setting (#26) --- ## Install in Seconds **From VS Code:** Press `Ctrl+P` (`Cmd+P` on Mac), then paste: ``` ext install DeepMyst.mysti ``` **Or** [install from the VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=DeepMyst.mysti) --- ## Choose Your AI Mysti works with the AI coding tools you already have. **No extra subscriptions needed.**

Agent Switching

| Provider | Best For | |----------|----------| | **Claude Code** | Deep reasoning, complex refactoring, thorough analysis | | **Codex** | Quick iterations, familiar OpenAI style | | **Gemini** | Fast responses, Google ecosystem integration | | **GitHub Copilot** | Multi-model access (Claude, GPT-5, Gemini) via GitHub subscription | | **Cline** | Plan/Act mode, structured task completion | | **Cursor** | Auto model selection, multi-model with Claude, GPT-5, Gemini | | **OpenClaw** | Real-time WebSocket streaming, configurable thinking levels | | **OpenCode** | Multi-backend agent (Anthropic, OpenAI, Google, Groq) | | **Qwen Code** | Alibaba's AI coding agent, deep reasoning | | **Ollama** | Local LLM inference, privacy-first, no subscription | | **LocalAI** | Self-hosted AI models, full control | **Switch providers with one click. No lock-in.** ### Why Mysti? | vs Copilot/Cursor | Mysti Advantage | |-------------------|-----------------| | Single AI | **Multi-agent brainstorming** — two AIs collaborate with 5 strategies | | Locked to one provider | **12 providers** — Claude, Codex, Gemini, Copilot, Cline, Cursor, OpenClaw, Manus, OpenCode, Qwen, Ollama, LocalAI | | Black box | **Full permission control** — read-only to full-access | | Generic responses | **16 personas** — architect, debugger, security expert... | | Manual workflow | **Autonomous mode** — AI works independently with safety controls | | No cross-agent routing | **@-mentions** — route tasks to specific agents inline | --- ## See It In Action

Mysti Chat Interface

Beautiful, modern chat interface with syntax highlighting, markdown support, and mermaid diagrams

Task List Rendering

Real-time task list rendering and progress tracking

--- ## Brainstorm Mode **Want a second opinion?** Enable Brainstorm Mode and let two AI agents tackle your problem together. **Choose any 2 of 12 agents** from the settings panel.

Brainstorm Mode

### 5 Collaboration Strategies | Strategy | Roles | Best For | |----------|-------|----------| | **Quick** | Direct synthesis | Simple tasks, fast answers | | **Debate** | Critic vs Defender | Architecture decisions, trade-offs | | **Red-Team** | Proposer vs Challenger | Security reviews, edge case discovery | | **Perspectives** | Risk Analyst vs Innovator | Greenfield design, technology selection | | **Delphi** | Facilitator vs Refiner | Complex problems, reaching consensus | ### Why Two AIs Beat One **Claude Code** (Anthropic), **Codex** (OpenAI), **Gemini** (Google), **GitHub Copilot**, **Cline**, **Cursor**, **OpenClaw**, **OpenCode**, **Qwen Code** (Alibaba), **Ollama**, and **LocalAI** have different training, different strengths, and different blind spots. When any two work together: - Each AI catches edge cases the other might miss - Different perspectives lead to more robust solutions - **Together** they debate, challenge each other, and synthesize the best solution It's like having a senior dev and a tech lead review your code—except they actually discuss it first. ### Convergence Detection During discussions, Mysti tracks agent agreement and position stability. When **auto-convergence** is enabled, the discussion exits early once agents reach consensus — saving time without sacrificing quality. ### Choose Your Team Configure which two agents collaborate in the **Settings Panel**:

Brainstorm Model Selection

| Combination | Best For | |-------------|----------| | Claude + Codex | Deep analysis meets rapid iteration | | Claude + Gemini | Thorough reasoning with fast validation | | Claude + Copilot | Compare native Claude vs Copilot's multi-model approach | | Cursor + Gemini | Multi-model flexibility with Google integration | | OpenClaw + Claude | WebSocket streaming with deep reasoning | | Qwen + Claude | Compare Alibaba and Anthropic reasoning | | OpenCode + Gemini | Multi-backend flexibility with Google speed | | Ollama + Claude | Local privacy meets cloud intelligence | [Full Brainstorm documentation](docs/BRAINSTORM.md) ### Intelligent Plan Detection When the AI presents multiple implementation approaches, Mysti automatically detects them and lets you choose your preferred path.

Plan Suggestions

*Requires at least 2 CLI tools installed. See [Requirements](#requirements).* --- ## Key Features ### Autonomous Mode Let the AI work independently with configurable safety controls: - **Safety Classifier**: Three levels — safe (auto-approve), caution (mode-dependent), blocked (always deny) - **Three Safety Modes**: Conservative, Balanced, Aggressive - **Learning Memory**: Remembers your permission preferences and improves over time - **Continuation Modes**: Goal-based or task-queue for extended autonomous sessions - **Audit Trail**: Every autonomous decision is logged for review

Selecting Autonomy Mode

[Full Autonomous Mode documentation](docs/AUTONOMOUS-MODE.md) ### @-Mention System Route tasks to specific agents and reference files inline:

@-Mention Tagging

``` @claude Review this code for security issues @src/auth.ts @gemini Suggest performance improvements for this file @claude Write tests, then @codex optimize them ``` - **File mentions**: `@filename` adds transient context - **Agent mentions**: `@agent` routes tasks to that provider - **Chaining**: Later agents receive earlier agents' responses as context [Full @-Mention documentation](docs/MENTIONS.md) ### Context Compaction Smart conversation management that prevents context overflow: - **Automatic**: Triggers when token usage approaches the threshold (default 75%) - **Native support**: Claude Code uses built-in `/compact` command - **Client-side**: Other providers use intelligent message summarization - **Per-panel tracking**: Each chat panel tracks usage independently [Full Compaction documentation](docs/COMPACTION.md) ### 16 Developer Personas Shape how your AI thinks. Select from specialized personas that change the AI's approach to your problems.

Personas and Skills Panel

| Persona | Focus | |---------|-------| | **Architect** | System design, scalability, clean structure | | **Debugger** | Root cause analysis, bug fixing | | **Security-Minded** | Vulnerabilities, threat modeling | | **Performance Tuner** | Optimization, profiling, latency | | **Prototyper** | Quick iteration, PoCs | | **Refactorer** | Code quality, maintainability | | + 10 more... | Full-Stack, DevOps, Mentor, Designer... | [Full Personas & Skills documentation](docs/PERSONAS-AND-SKILLS.md) --- ### Quick Persona Selection Select personas directly from the toolbar without opening panels.

Toolbar Persona Selection

--- ### Smart Auto-Suggestions Mysti automatically suggests relevant personas and actions based on your message.

Auto Suggestions

--- ### Conversation History Never lose your work. All conversations are saved and easily accessible.

Conversation History

--- ### Quick Actions on Welcome Get started fast with one-click actions for common tasks.

Quick Actions

--- ### Extensive Settings Fine-tune every aspect of Mysti including token budgets, access levels, and brainstorm mode.

Settings Panel

--- ## Requirements **Already paying for Claude, ChatGPT, Gemini, or GitHub Copilot? You're ready to go.** Mysti works with your existing subscriptions—no additional costs! | CLI Tool | Subscription | Install | |----------|--------------|---------| | **Claude Code** (recommended) | Anthropic API or Claude Pro/Max | `npm install -g @anthropic-ai/claude-code` | | **GitHub Copilot CLI** | GitHub Copilot Pro/Pro+/Business | `npm install -g @github/copilot-cli` | | **Gemini CLI** | Google AI API or Gemini Advanced | `npm install -g @google/gemini-cli` | | **Codex CLI** | OpenAI API | Follow OpenAI's installation guide | | **Cline** | Depends on model provider | `npm install -g cline` | | **Cursor** | Cursor subscription | `curl https://cursor.com/install -fsS \| bash` | | **OpenClaw** | OpenClaw account | `npm install -g openclaw@latest && openclaw onboard --install-daemon` | | **OpenCode** | Provider API keys (Anthropic, OpenAI, etc.) | `npm i -g opencode-ai@latest` | | **Qwen Code** | Qwen OAuth or API keys | `npm install -g @qwen-code/qwen-code@latest` | | **Ollama** | Local (no subscription needed) | [Install from ollama.com](https://ollama.com) | | **LocalAI** | Local (no subscription needed) | [Install from localai.io](https://localai.io) | You only need **one** CLI to get started. Install **any two** to unlock Brainstorm Mode. --- ## Quick Start ### 1. Install Mysti **Option A:** Press `Ctrl+P` (`Cmd+P` on Mac), paste and run: ``` ext install DeepMyst.mysti ``` **Option B:** [Install from VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=DeepMyst.mysti) ### 2. Install a CLI Tool ```bash # Claude Code (recommended) npm install -g @anthropic-ai/claude-code claude auth login # Or GitHub Copilot CLI (access Claude, GPT-5, Gemini via GitHub) npm install -g @github/copilot-cli copilot # then use /login command # Or Gemini CLI npm install -g @google/gemini-cli gemini auth login # Or Cursor curl https://cursor.com/install -fsS | bash agent login # Or OpenClaw npm install -g openclaw@latest && openclaw onboard --install-daemon openclaw login # Or OpenCode npm i -g opencode-ai@latest opencode auth login # Or Qwen Code npm install -g @qwen-code/qwen-code@latest qwen # then type /auth ``` For Brainstorm Mode, install any two CLI tools. ### 3. Open Mysti - Click the **Mysti icon** in the Activity Bar, or - Press `Ctrl+Shift+M` (`Cmd+Shift+M` on Mac) ### 4. Start Coding Type your request and let the AI assist you! --- ## Slash Commands Access skills and actions quickly with the built-in slash command menu.

--- ## 12 Toggleable Skills Mix and match behavioral modifiers: - **Concise** - Clear, brief communication - **Test-Driven** - Tests alongside code - **Auto-Commit** - Incremental commits - **First Principles** - Fundamental reasoning - **Scope Discipline** - Stay focused on the task - And 7 more... [Full Personas & Skills documentation](docs/PERSONAS-AND-SKILLS.md) --- ## Permission Controls Stay in control of what the AI can do: - **Read-only** - AI can only read, never modify - **Ask-permission** - Approve each file change - **Full-access** - Let the AI work autonomously

Permission Controls Demo

--- ## Configuration ### Essential Settings ```json { "mysti.defaultProvider": "claude-code", "mysti.brainstorm.agents": ["claude-code", "google-gemini"], "mysti.brainstorm.strategy": "quick", "mysti.accessLevel": "ask-permission" } ``` ### Provider Settings | Setting | Default | Description | |---------|---------|-------------| | `mysti.defaultProvider` | `claude-code` | Primary AI provider | | `mysti.claudePath` | `claude` | Path to Claude CLI | | `mysti.codexPath` | `codex` | Path to Codex CLI | | `mysti.geminiPath` | `gemini` | Path to Gemini CLI | | `mysti.copilotPath` | `copilot` | Path to Copilot CLI | | `mysti.clinePath` | `cline` | Path to Cline CLI | | `mysti.cursorPath` | `agent` | Path to Cursor CLI | | `mysti.openclawPath` | `openclaw` | Path to OpenClaw CLI | | `mysti.opencodePath` | `opencode` | Path to OpenCode CLI | | `mysti.qwenCodePath` | `qwen` | Path to Qwen Code CLI | | `mysti.ollamaPath` | `ollama` | Path to Ollama CLI | | `mysti.localaiPath` | `localai` | Path to LocalAI CLI | ### Brainstorm Settings | Setting | Default | Description | |---------|---------|-------------| | `mysti.brainstorm.agents` | `["claude-code", "openai-codex"]` | Which 2 agents to use | | `mysti.brainstorm.strategy` | `quick` | Strategy: `quick`, `debate`, `red-team`, `perspectives`, `delphi` | | `mysti.brainstorm.autoConverge` | `true` | Auto-exit when agents converge | | `mysti.brainstorm.maxDiscussionRounds` | `3` | Maximum discussion rounds | ### Autonomous Settings | Setting | Default | Description | |---------|---------|-------------| | `mysti.autonomous.safetyMode` | `balanced` | `conservative`, `balanced`, `aggressive` | | `mysti.autonomous.blockPatterns` | `[]` | Custom patterns to always block | ### Compaction Settings | Setting | Default | Description | |---------|---------|-------------| | `mysti.compaction.enabled` | `true` | Enable context compaction | | `mysti.compaction.threshold` | `75` | Compaction threshold (% of context window) | ### General Settings | Setting | Default | Description | |---------|---------|-------------| | `mysti.accessLevel` | `ask-permission` | File access level | | `mysti.agents.autoSuggest` | `true` | Auto-suggest personas | | `mysti.agents.maxTokenBudget` | `0` | Max tokens for agent context (0 = unlimited) | [Full Provider documentation](docs/PROVIDERS.md) --- ## Keyboard Shortcuts | Action | Windows/Linux | Mac | |--------|---------------|-----| | Open Mysti | `Ctrl+Shift+M` | `Cmd+Shift+M` | | Open in New Tab | `Ctrl+Shift+N` | `Cmd+Shift+N` | --- ## Commands | Command | Description | |---------|-------------| | `Mysti: Open Chat` | Open the chat sidebar | | `Mysti: New Conversation` | Start fresh | | `Mysti: Add to Context` | Add file/selection to context | | `Mysti: Clear Context` | Clear all context | | `Mysti: Open in New Tab` | Open chat as editor tab | --- ## Documentation | Guide | Description | |-------|-------------| | [Providers](docs/PROVIDERS.md) | All 12 providers — setup, models, features | | [Brainstorm Mode](docs/BRAINSTORM.md) | 5 strategies, convergence, team selection | | [Personas & Skills](docs/PERSONAS-AND-SKILLS.md) | 16 personas, 12 skills, custom agents | | [Autonomous Mode](docs/AUTONOMOUS-MODE.md) | Safety system, memory, continuation modes | | [@-Mentions](docs/MENTIONS.md) | Agent routing and file context | | [Compaction](docs/COMPACTION.md) | Context management and summarization | | [Architecture](docs/ARCHITECTURE.md) | Technical internals and extension points | | [Features](docs/FEATURES.md) | Complete feature reference | --- ## Telemetry Mysti collects **anonymous** usage data to improve the extension: - Feature usage patterns - Error rates - Provider preferences **No code, file paths, or personal data is ever collected.** Respects VSCode's telemetry setting. Disable via: Settings > Telemetry: Telemetry Level > off --- ## Contributors Thanks to everyone who has helped make Mysti better!

Want to join them? Check out the [Contributing](#contributing) section below. --- ## Star History If Mysti has been useful to you, consider giving it a star — it helps others discover the project and keeps us motivated!

--- ## Contributing We welcome contributions! Whether it's bug reports, feature requests, or code contributions. - **Good First Issues**: Look for [`good first issue`](https://github.com/DeepMyst/Mysti/labels/good%20first%20issue) labels - **Development**: Press `F5` in VS Code to launch Extension Development Host - **Pull Requests**: Fork, create a feature branch, and submit a PR See [CONTRIBUTING.md](CONTRIBUTING.md) for detailed guidelines. --- ## License Apache License 2.0 — free to use, modify, and distribute, including for commercial purposes. See the `LICENSE` file for full text. ---

Install • Report Issue • GitHub

Mysti — Built by DeepMyst Inc
_{Made with Mysti}