Shofer is the open-source agentic coding platform.
Take full control of your data, your models, and how your tools work — with parallel tasks, custom modes, and deep codebase intelligence.
Website · Docs · Install · GitHub · Roo-Code Migration · Copilot Migration
50+ tools. Custom modes. Parallel tasks. Semantic code search. Open source under Apache 2.0.
- Open Shofer — Click the Shofer icon in the Activity Bar, or press the assigned keyboard shortcut
- Configure a Provider — Open Settings → Providers → New Profile. Shofer works with Anthropic, OpenAI, Google Gemini, DeepSeek, OpenRouter, Ollama (local), and more. Your API keys are stored securely in your OS credential store.
- Start a Task — Type a message in the chat input bar. Shofer reads your workspace, searches code, runs commands, and edits files. Review every change in the File Changes Panel — accept, revert, or diff each one.
Try the built-in /init slash command to analyze your codebase and create an AGENTS.md file with build commands, code style, and project conventions.
| UI Element | Purpose |
|---|---|
| Chat View | Interact with the AI — type messages, see results |
| Task Selector | Switch between multiple parallel tasks in a tree hierarchy |
| Mode Selector | Choose Code, Architect, Ask, Debug, Orchestrator, or custom modes |
| API Config Selector | Pick which AI provider and model to use per task |
| Worktree Selector | Create and select git worktrees for isolated parallel work |
| File Changes Panel | Review, accept, revert, or diff every file Shofer modifies |
| Context Window Bar | Monitor token usage and cost for the current task |
Shofer ships with 5 built-in modes — choose from the Mode Selector dropdown in the chat input bar:
| Mode | Best For |
|---|---|
| 💻 Code | Writing, modifying, and refactoring code. Broadest tool access. |
| 🏗️ Architect | Planning and designing before writing code. Read + markdown-only writes. |
| ❓ Ask | Getting explanations, answers, or recommendations. Read-only + MCP. |
| 🪲 Debug | Troubleshooting errors and diagnosing root causes. |
| 🪃 Orchestrator | Coordinating complex multi-step work by delegating to sub-tasks. |
Create your own modes (Reviewer, Search, Opinion, Browser, and more) via .shofer/shofermodes files at the project or global level. Control exactly which tool categories are available per mode, with file-scoped restrictions.
Learn more: User Manual • Custom Modes
Shofer ships with two built-in multi-agent workflows — formal, deterministic coordination of multiple AI agents specified in .slang files:
| Workflow | Description |
|---|---|
| 🪲 Collaborative Debug | Two developers triage independently, converge on root cause, one fixes while the other reviews |
| 🔧 Implement a Feature | Architect designs → you approve → Developer implements → Reviewer inspects — iterate until done |
Launch a workflow via New… → New Workflow in the chat toolbar. Create your own custom workflows as .slang files under .shofer/workflows/ (project) or ~/.shofer/workflows/ (global).
Learn more: User Manual → Workflows • Built-in Workflows SoT • Slang Language Spec
Shofer supports true parallel tasks organized in a tree hierarchy. Start multiple conversations simultaneously — each runs independently with its own mode, provider, and context.
- Background subtasks — fan out work without blocking the parent. The Orchestrator mode delegates to specialized sub-tasks automatically.
- Message queuing — type ahead while the AI is working. Click Send Now to redirect it immediately.
- Task states — colored dots show each task's status: idle, running, waiting for input, waiting on subtask, completed, or errored.
Run parallel tasks on different branches — all in one VS Code window:
- Worktrees live at
.shofer/worktrees/<name>/inside your workspace - Create, switch, and delete worktrees from the Worktree Selector in the Task Header
- Each task can be scoped to a specific worktree for branch-isolated parallel work
- No more stash/commit gymnastics or multiple VS Code windows for PRs
Read the full worktree documentation
Build a semantic search index of your codebase and git history so the AI can find code and commits by meaning — not just keywords.
| Index | What It Searches | Tool |
|---|---|---|
| Code | Functions, classes, types, comments | rag_search |
| Git History | Commit messages and metadata | git_search |
Requires a reachable Qdrant v1.14.x server (local or remote). Configure in Settings → RAG Indexing — pick an embedding provider, point to your Qdrant instance, and click Start Indexing. A file watcher keeps the index up to date as you edit.
Read the full RAG indexing documentation
The Assistant Agent is a persistent, read-only AI companion that accumulates codebase knowledge over time — surviving task completion and VS Code restarts.
- Runs on a low-cost model with a large context window (you choose the model)
- Answers questions from any task via the
ask_assistant_agenttool - Strictly read-only — can read files, search code, and look up symbols; cannot write or execute
- KV-cache friendly — append-only context window keeps provider costs minimal
- Learns organically — each question adds context, building an ever-richer understanding of your codebase
Enable in Settings and choose a lightweight model (e.g., Gemini Flash, GPT-4o-mini).
Read the full Assistant Agent documentation
Shofer recognizes several files in your workspace that control its behavior. These files are write-protected — Shofer cannot modify them without your explicit approval.
| File / Directory | Purpose |
|---|---|
.shofer/shoferignore |
Hide files from Shofer (same syntax as .gitignore) |
.shofer/shofermodes |
Define custom AI modes for your project |
AGENTS.md |
Project rules injected into every task |
.shofer/rules/ |
Mode-agnostic rules (always active) |
.shofer/rules-<mode>/ |
Rules for a specific mode (e.g., rules-code/) |
.shofer/commands/ |
Custom slash commands |
.shofer/skills/ |
Domain-specific skill instructions |
.shofer/mcp.json |
Per-project MCP server configuration |
| Section | What You Can Do |
|---|---|
| Providers | Add API profiles, switch models, configure endpoints |
| Modes | Create and edit custom modes with per-tool access control |
| Auto-Approval | Toggle which tool categories run without asking permission |
| MCP Servers | Connect external tools (browser, databases, Kubernetes) |
| RAG Indexing | Build a semantic search index of your codebase |
| Context Management | Control when conversations get condensed |
| Limits | Set per-task USD cost caps and API request limits |
Key recommendations: start with auto-approval toggles OFF and enable incrementally. Set shofer.defaultCostLimit to cap spending. Configure a command allowlist for shell operations.
Coming from Roo-Code? Shofer is a major architectural improvement with parallel tasks, async MCP calling, semantic code & git log indexing, native worktree support, and background subtasks. Run /migrate-from-roocode to migrate your configuration. Read the migration guide →
Coming from GitHub Copilot? Shofer gives you full model choice, local-first privacy, 50+ native tools, and parallel task execution. Run /migrate-from-copilot to migrate your configuration. Read the migration guide →
- Discord — Chat with the team, get help, share feedback
- Reddit — Community discussions and tips
- GitHub Issues — Bug reports and tracking
Shofer is open source (Apache 2.0). Contributions are welcome — read CONTRIBUTING.md and check the roadmap.
- User Manual: The complete guide to every feature, setting, and concept in Shofer.
- Developer Documentation: Official docs for installing, configuring, and mastering Shofer.
- GitHub Issues and Feature Requests: Report bugs, feature requests, and track development.
- Clone the repo:
git clone https://github.com/shofer-dev/shofer.git- Install dependencies:
pnpm install- Run the extension:
There are several ways to run the Shofer extension:
For active development, use VSCode's built-in debugging:
Press F5 (or go to Run → Start Debugging) in VSCode. This will open a new VSCode window with the Shofer extension running.
- Changes to the webview will appear immediately.
- Changes to the core extension will also hot reload automatically.
To build and install the extension as a VSIX package directly into VSCode:
pnpm install:vsix [-y] [--editor=<command>]This command will:
- Ask which editor command to use (code/cursor/code-insiders) - defaults to 'code'
- Uninstall any existing version of the extension.
- Build the latest VSIX package.
- Install the newly built VSIX.
- Prompt you to restart VS Code for changes to take effect.
Options:
-y: Skip all confirmation prompts and use defaults--editor=<command>: Specify the editor command (e.g.,--editor=cursoror--editor=code-insiders)
If you prefer to install the VSIX package manually:
- First, build the VSIX package:
pnpm vsix
- A
.vsixfile will be generated in thebin/directory (e.g.,bin/shofer-<version>.vsix). - Install it manually using the VSCode CLI:
code --install-extension bin/shofer-<version>.vsix
We use changesets for versioning and publishing. Check our CHANGELOG.md for release notes.
Please note that Shofer, Inc does not make any representations or warranties regarding any code, models, or other tools provided or made available in connection with Shofer, any associated third-party tools, or any resulting outputs. You assume all risks associated with the use of any such tools or outputs; such tools are provided on an "AS IS" and "AS AVAILABLE" basis. Such risks may include, without limitation, intellectual property infringement, cyber vulnerabilities or attacks, bias, inaccuracies, errors, defects, viruses, downtime, property loss or damage, and/or personal injury. You are solely responsible for your use of any such tools or outputs (including, without limitation, the legality, appropriateness, and results thereof).
We love community contributions! Get started by reading our CONTRIBUTING.md.