OpenCoworker is an open-source AI agent application that allows users to assign real-world tasks to AI models that can operate on files, documents, coding projects, and connected tools.
Unlike traditional AI chat interfaces, OpenCoworker can:
Work inside local folders
Create and modify files
Schedule recurring tasks
Connect to external tools
Use MCP servers
Operate with multiple AI providers
Maintain long-running workflows
The platform is designed to function more like a digital coworker than a conversational assistant.
The result is a platform that attempts to bridge the gap between AI chatbots and autonomous agents by allowing AI models to work directly with files, folders, automations, and external tools on a user's machine.
Relationship to aisuite
A common misconception is that OpenCoworker and aisuite are the same product.
In reality, aisuite is a lightweight open-source library created to provide a unified API across multiple AI providers, allowing developers to switch between models using a consistent interface. It supports providers such as OpenAI, Anthropic, Google, Ollama, AWS, Mistral, and others.
OpenCoworker builds on similar multi-model concepts but focuses on practical task execution rather than API abstraction.
For developers, aisuite serves as infrastructure.
For end users, OpenCoworker serves as the application layer.
Download aisuite OpenWorker v0.3.0 - Software Mirrors |
|---|
aisuite OpenWorker v0.3.0 for WindowsOpenWorker_0.3.0_x64_en-US.msi | 73.51 MB OpenWorker_0.3.0_x64-setup.exe | 61.75 MB OpenWorker-windows.msi | 73.51 MB OpenWorker-windows-setup.exe | 61.75 MB |
aisuite OpenWorker v0.3.0 for macOSOpenWorker_0.3.0_x86_64.dmg | 79.16 MB OpenWorker_0.3.0_aarch64.dmg | 77.47 MB OpenWorker-macos-x64.dmg | 79.16 MB OpenWorker-macos-x64.app.tar.gz | 79.98 MB OpenWorker-macos-arm64.dmg | 77.47 MB OpenWorker-macos-arm64.app.tar.gz | 78.62 MB |
Others Download related to aisuite OpenWorker v0.3.0openworker-linux-x86_64.tar.gz | 91.95 MB openworker-linux-aarch64.tar.gz | 90.19 MB openworker-0.3.0-linux-x86_64.tar.gz | 91.95 MB openworker-0.3.0-linux-aarch64.tar.gz | 90.19 MB |
aisuite OpenWorker v0.3.0 Source Code
aisuite OpenWorker v0.3.0 Source code (zip)
aisuite OpenWorker v0.3.0 Source code (tar.gz)
|
aisuite OpenWorker v0.3.0 Release Notes: OpenWorker 0.3.0
Agent Security OpenWorker now provides multiple layers of protection for AI agents. The approval system and the permission ladder decide what an agent _may_ do, and a secure runtime now sets a hard limit on what it _can_ do. OpenWorker integrates the OpenShell secure runtime from the NVIDIA Open Agent Safety Platform, announced this week. It is an enforceable boundary outside the model and the agent harness. Every agent runs in its own Linux container, with Landlock and seccomp on every process. This enables Users to restrict AI agents to only allowed files and allowed websites. Even if an agent is misled by a prompt injection or drifts from its task, it cannot read your other files, take your keys, or reach anywhere else. Docker (or Docker Desktop) needs to be installed on Mac, Windows or Linux to allow OpenWorker to setup and use OpenShell. In absence of Docker, sandboxing is still provided using OS specific primitives. Detailed architecture to come soon as a blog and docs. Thanks to @devikaverma for this researching and enabling above techniques inside OpenWorker to provide a more secure way to run AI agents.
Set it up in Settings ▸ Sandbox.
Machines. Run OpenWorker on a different machine and use it from the desktop app. Sessions live on that remote machine and keep running when your laptop is closed. Join a machine from Settings ▸ Machines.
This enables several use cases like 24x7 code reviewer agent, or an agent that answers question from Slack.
openworker command line. On Linux, install it with one command:
curl -fsSL https://raw.githubusercontent.com/andrewyng/openworker/main/packaging/install.sh | sh
Or download it below: openworker-linux-x86_64.tar.gz or openworker-linux-aarch64.tar.gz.
The command line can be used to make a remote openworker join the Desktop app, allowing the user to control it remotely.
Also: the app is in English and Simplified Chinese; many fixes to the agent loop, approvals, scheduled tasks and MCP tools.
More languages are in the roadmap.
Downloads: Mac (Apple Silicon: OpenWorker-macos-arm64.dmg; Intel: OpenWorker-macos-x64.dmg), Windows (OpenWorker-windows-setup.exe), Linux (above).
What's Changed- Updating README by @rohitprasad15 in #569
- security(teams): enforce worker visibility on board item reads by @rakeshutekar in #585
- fix(teams): enforce attachment read authorization by @rakeshutekar in #586
- Add GUI internationalization with English and Simplified Chinese by @jasmine889966 in #127
- fix: scheduled task can run twice when an approval lands on a tick by @Saidheerajgollu in #379
- Parse inbox reply intent from the leading word, not substrings by @Saidheerajgollu in #24
- Reject path-traversal session ids in the conversation store by @Saidheerajgollu in #55
- Recover truncated tool calls, and never pass a leaked one off as an answer by @hacksics in #219
- Tolerate a corrupt line when loading a conversation .jsonl by @Saidheerajgollu in #56
- Make the conversation-log shrink rewrite atomic (prevent history loss on a mid-write crash) by @lifrary in #70
- fix: repair tool-call/result pairing on load to prevent unrecoverable 400 errors by @rkfshakti in #350
- fix(skills): confine staged upload tokens by @mo-tunn in #548
- MCP permission model: EXTERNAL floor, durable per-tool trust, fixed approval cards (OPE-136) by @devikaverma in #598
- fix(engine): agent loop fixes for long, tool-heavy sessions (OPE-156) by @devikaverma in #669
- setup_dev_env.sh: install the bedrock extra so a fresh env passes the test suite by @xiaonancui in #285
- fix: handle Windows drive-letter paths in grep (ripgrep) output by @engmohamedsalah in #123
- fix: preserve exception chain in RelayHub.wait_dispatched by @Anuj04432 in #172
- fix: add httpx2 dev dependency to silence Starlette TestClient deprec… by @Aadhithya-arulvanan in #183
- security: reject shell variable expansion in read-only grants by @harneet2512 in #566
- fix: make Bedrock verification and tests tolerate missing optional boto3 by @tyoon10 in #554
- fix(tests): import ExceptionGroup from backport for Python 3.10 compatibility by @tyoon10 in #555
- Escape LIKE wildcards when re-keying board cursors by @tyoon10 in #556
- fix: add coverage measurement and reporting to CI by @rkfshakti in #328
- Remote machines, agent-team cards, and connectors across machines by @rohitprasad15 in #672
- CLI:
openworker join, up and machine commands; package publishes as openworker by @rohitprasad15 in #673 - Add Team View and harden agent-team coordination by @rohitprasad15 in #679
- Add support for Sandboxing using NVIDIA OpenShell by @devikaverma in #684
- Add docs for OpenShell support by @rohitprasad15 in #685
- feat(sandbox): guided OpenShell setup, readiness checks, switching for open sessions (OPE-205, OPE-206, OPE-207, OPE-208, OPE-209) by @devikaverma in #699
- Sandboxing improvements for Windows and Mac by @rohitprasad15 in #700
- Simplify the Sandbox settings page. by @rohitprasad15 in #701
- Fix KeyError 'none' crash in openai_compat_effort for unlisted model ids by @aashish254 in #677
- OpenShell clean-up removes only the sandboxes its own state folder made by @rohitprasad15 in #704
- Linux: build openworker as one program, no Python needed by @rohitprasad15 in #705
- fix(mentions): a corrupt mention_threads.json must not brick server startup by @Lesereingrape in #690
- fix(providers): handle effort "none" on every provider, Anthropic omits it, OpenAI-compat sends it (#702) by @devikaverma in #703
- fix(security): refuse find's fprint family under a bare allowlist entry by @rkfshakti in #688
- Release 0.3.0: Linux download and install script by @rohitprasad15 in #706
New Contributors- @rakeshutekar made their first contribution in #585
- @jasmine889966 made their first contribution in #127
- @Saidheerajgollu made their first contribution in #379
- @hacksics made their first contribution in #219
- @lifrary made their first contribution in #70
- @mo-tunn made their first contribution in #548
- @xiaonancui made their first contribution in #285
- @engmohamedsalah made their first contribution in #123
- @Anuj04432 made their first contribution in #172
- @Aadhithya-arulvanan made their first contribution in #183
- @harneet2512 made their first contribution in #566
- @tyoon10 made their first contribution in #554
- @aashish254 made their first contribution in #677
- @Lesereingrape made their first contribution in #690
Full Changelog: v0.2.1...v0.3.0
|
Key Features of OpenCoworker
Bring Your Own Model
One of OpenCoworker's strongest features is model flexibility.
Users can connect:
OpenAI models
Anthropic Claude models
Google Gemini models
Local Ollama models
Switching between models can be done per conversation or task.
Real File Access
Unlike chatbot interfaces that generate text outputs, OpenCoworker can work directly with user files.
The agent can:
Everything remains accessible as standard files on the user's system.
Automations
The application supports scheduled workflows.
Examples include:
These tasks can execute automatically without requiring continuous user supervision.
MCP Support
OpenCoworker supports the Model Context Protocol (MCP), allowing integration with external tools and services through standardized interfaces.
This significantly expands the capabilities of the platform beyond simple file manipulation.
Security Controls
The application includes approval-based workflows for potentially risky operations.
Users remain in control of sensitive actions rather than granting unrestricted autonomy to AI agents.
User Experience
The experience feels noticeably different from ChatGPT, Claude, or Gemini.
Instead of repeatedly prompting a chatbot, users assign objectives and review outputs.
For example, a user might ask OpenCoworker to:
Research a topic
Generate a report
Create project files
Update documentation
Organize data
The agent then performs the work and produces actual files rather than simply displaying responses in a chat window.
The interface is relatively approachable compared to many agent frameworks, though it still requires a basic understanding of AI models, permissions, and workflows.
Performance
Performance depends primarily on:
Simple document-generation tasks are typically fast.
More complex workflows involving coding, MCP tools, or large projects can take significantly longer, depending on the chosen model and environment.
Open Source Advantages
Being open source offers several benefits:
Transparency
Community contributions
Self-hosting options
No vendor lock-in
Greater privacy control
Users can inspect the codebase and customize deployments according to their needs.
Community Feedback
Developer discussions generally highlight two major strengths:
Simplicity
Model flexibility
Many users appreciate the philosophy inherited from aisuite, where switching AI providers requires minimal effort. Others view unified AI abstraction layers as useful for experimentation but question how often model switching is required in production.
Community feedback regarding aisuite itself frequently praises its beginner-friendly design and low learning curve, though some developers note that larger frameworks may provide more extensive documentation and ecosystem support.
Limitations
Despite its promise, OpenCoworker remains an early-stage project.
Current limitations include:
Smaller community than established AI platforms
Reliance on external model APIs
Requires API keys or local models
Limited ecosystem compared to mature agent frameworks
Some workflows still require manual supervision
The platform is better viewed as a productivity assistant than a fully autonomous employee.
Pros
Open source
Supports multiple AI providers
Real file creation and editing
MCP integration
Automation capabilities
User-controlled permissions
Flexible deployment options
Beginner-friendly compared to many agent frameworks
Cons
Early-stage project
Smaller ecosystem
Requires external AI providers or local models
Advanced workflows may need technical knowledge
Limited long-term production track record
Who Should Use OpenCoworker?
OpenCoworker is ideal for:
It is particularly useful for people who want AI to generate actual work products rather than simply provide conversational responses.
OpenCoworker offers an interesting vision of what desktop AI agents can become. By combining real file access, automation, multi-model support, and open-source flexibility, it delivers capabilities that go beyond traditional AI chat applications. While the platform is still young and its ecosystem is evolving, it already provides a compelling environment for users looking to experiment with practical AI coworkers rather than simple conversational assistants.
