Give it a goal. It plans, splits, and runs the work in parallel — then hands you the result.

Not a chat box — a desktop AI team that actually does the work.

Desktop Native16x Parallel Execution100% Local Privacy
Download NowWatch Demo

Orchestrator + parallel subagents · Skills · MCP · runs locally

Works with all major models — use your own API keys

  • OpenAI
  • Anthropic
  • Google
  • Meta
  • GLM
  • DeepSeek

What would you like to do today?

Free

Type your question or pick a suggestion below

Refactor Legacy Code

Drop in a decade-old project. Doforu analyzes the architecture, breaks it down, and refactors module by module in parallel.

See how it works

Deep Research

Need market intel or technical research? Multiple Agents search, cross-reference, and synthesize findings simultaneously.

See how it works

Full-stack Development

Give it a goal — the orchestrator dispatches dedicated Agents for frontend, backend, API and database, then delivers it all at once.

See how it works

16x

Parallel Execution

15+

Model Support

3

Platforms

100%

Local Privacy

Core Capabilities

Orchestrator · Plan · Agent Desktop AI agent orchestrator. Give it a goal — your AI team works in parallel, zero-wait, no context loss.

Orchestrator Intelligent Dispatch

Receives your instruction, understands intent, evaluates complexity, breaks tasks into parallel subtasks, and dynamically assigns them to the most suitable subagents.

Plan Instant Planning

Analyzes your requirements and outputs a complete execution blueprint (steps, tool selection, estimated time). Confirm manually, then switch to Agent or Orchestrator mode to run.

Agent Parallel Execution

Multiple subagents work simultaneously, each responsible for their part, automatically correcting errors, collaborating, and aggregating results into a complete deliverable.

100% Local Privacy

All code, data, and API calls stay on your machine. No cloud upload, no data leakage — your sensitive work never leaves your desktop.

Evolving Skills

Create or refine custom Skills in one sentence. Teach it once, reuse forever. The more your team uses it, the better it understands you.

MCP Ecosystem Integration

Seamlessly connects with databases, Figma, Slack, GitHub, and other external tools to extend Agent capabilities.

Desktop-Native Experience

Direct read/write to the local file system, deeply integrated with your development environment. Supports 15+ local and international models.

30-Day Guarantee

Not satisfied? Full refund within 30 days. No questions asked.

100% Local Privacy

Your code and API keys never leave your machine. No cloud upload.

1 Year Updates

All major updates included. Renew at $49/yr to keep receiving.

4h Response Support

Enterprise-grade support. We respond within 4 hours.

Pricing

Free to start. Pro removes run limits and adds parallel delivery & multi-window.

Free

Chat unlimited, 20 Agent/Plan/Orchestrator runs per day combined

$0
Get Started
  • Chat: Unlimited
  • Agent / Plan / Orchestrator: 20 runs/day combined
  • Windows: 1 window
  • Custom Skills: Up to 5
  • MCP servers: 3
  • macOS, Windows & Linux clients
  • 15+ model support

Pro

Most Popular

Perpetual license, use forever

$99$49

One-time purchase, lifetime use

Perpetual license. Includes 1 year of updates. Renew for $49/year to keep receiving updates.

  • Chat: Unlimited
  • Agent: Unlimited
  • Plan: Unlimited
  • Orchestrator: Unlimited
  • Subagents: Unlimited
  • Per-agent model selection
  • Windows: Up to 8 windows
  • Custom Skills: Unlimited
  • MCP servers: Unlimited
  • macOS, Windows & Linux clients
  • 15+ model support
  • Custom models: Supported
  • Perpetual license — use the software forever
  • 1 year of updates included
  • Proactive context compression

Enterprise

For teams and organizations

Custom
Contact Us
  • Everything in Pro
  • Priority support (4h response)
  • Custom integrations & deployment
  • Team workspaces & shared Skills
  • Dedicated account manager

Frequently Asked Questions

What is Doforu and how is it different from Cursor?
Doforu is a desktop AI agent orchestrator, unlike Cursor (IDE plugin) or Claude Code (terminal CLI). Give it a goal — your entire AI team delivers in parallel, zero-wait, no context loss. Supports Plan / Agent / Orchestrator modes.
Which AI models are supported?
Supports 15+ local and international models: OpenAI (GPT-6 Astra), Anthropic (Claude Opus 5.5), Google (Gemini 4 Argon), DeepSeek (V4.1), Alibaba Qwen (Qwen3.8-Max), ByteDance Doubao (Seed 2.1), Moonshot (Kimi K3), MiniMax (M3), Zhipu (GLM-5.3), and more. Also compatible with local models via Ollama and vLLM.
How do I create custom Skills?
Describe it in natural language within Doforu to create a Skill — teach it once, reuse forever. Skills support local directory management and Claude Code-compatible formats. Team sharing is coming soon.
What's the difference between Free and Pro?
Free includes unlimited Chat mode, 20 Plan/Agent/Orchestrator runs per day combined, up to 5 custom Skills, and 3 MCP servers. Pro unlocks unlimited runs, unlimited Skills, long-term Memory, unlimited MCP servers, and proactive context compression.
How are API costs calculated?
Doforu is a $49 one-time perpetual license — not a subscription, and not a per-run platform fee. API usage is billed directly by the model provider through your own API Key. Doforu does not intermediate, markup, or profit from API usage — costs are fully transparent.
Which platforms are supported?
macOS 12+, Windows 10+, and Linux are all supported. On Linux we ship AppImage, .deb, and .rpm packages. Download the client for your platform, select a model, and enter your API Key to get started.
How do I get help?
Visit the documentation center for self-service help, or submit feedback via [email protected]. Enterprise users get support with a 4-hour response time.
How does Doforu compare to AutoGPT or CrewAI?
Unlike AutoGPT (single-agent, loop-based) or CrewAI (multi-agent orchestration framework requiring Python coding), Doforu is a desktop application with a visual interface. It features Plan mode for structured task decomposition, Agent mode for autonomous execution, and Orchestrator mode for parallel agent dispatch—all without writing a single line of code.
Is Doforu really free? What are the limitations?
Yes, the Free tier is genuinely free with no time limit. It includes unlimited Chat mode, 20 combined runs per day for Plan/Agent/Orchestrator modes, up to 5 custom Skills, and 3 MCP server connections. The Pro tier removes these limits for power users who need more runs and advanced features like long-term Memory.
Can I use Doforu with my existing codebase?
Absolutely. Doforu works directly with your existing projects on your local machine. You can point it to any directory, and it will understand your codebase structure, read files, execute commands, and make changes using AI Agents—all while keeping your data local and secure.
What is the Orchestrator mode and how does it work?
Orchestrator mode is Doforu's flagship feature for complex tasks. It automatically breaks down your request into sub-tasks, spawns multiple AI Agents to work on them in parallel, and then synthesizes the results into a complete deliverable. Think of it as having a team of AI assistants working simultaneously on different parts of your problem.

Get Started in Seconds

Latest release: v3.7.3

Native macOS, Windows & Linux app. Your AI team, ready to work in parallel.

Download for Windows

Installer (.exe) for most Windows PCs

Download for macOS

For Macs with Intel chips

Other packages

Download for Linux

Portable AppImage, no install needed

Other packages