A desktop app that splits complex tasks — research, document creation, data analysis — across multiple AI agents specialized by role, and runs them in parallel. Built on CAMEL-AI’s multi-agent framework, it lets you choose where you connect, so you can operate without sending your data outside. Open source (Apache 2.0), available on Mac, Windows, and Linux. After its July 2025 release, it hit number one on GitHub Trending and saw rapid adoption in a short window.
Key Features
- Multi-agent workforce: Runs several agents specialized in research, coding, document editing, image/audio processing, and more in parallel, splitting a single complex task across them
- Local execution / BYOK: Connect via cloud API, your own API keys, local models, or an enterprise gateway — whichever fits, including setups that never send data outside
- Custom MCP server integration: Add Model Context Protocol-compliant external tools and services as connectors from the settings screen, then let your agents use them
- Scheduled execution: Schedule tasks to run on a recurring basis and keep the workforce running continuously
- Plugin / skill extensions: Pull in additional skills and plugins from the agent-plugins repository or the Skills directory
Pricing
| Plan | Monthly Price | Key Features |
|---|---|---|
| Free | Free | 500 credits at signup plus more for invites, BYOK/local model support, email support |
| Plus | $24.99 ($19.90/mo billed annually) | 2,000 task credits per month, priority email support, 7-day free trial |
| Pro | $129.99 ($99.99/mo billed annually) | 10,000 task credits per month, high-priority user support, 7-day free trial |
| Teams | Coming soon | Shared library, centralized billing, seat management |
| Enterprise | Custom | Local/self-hosted deployment, enterprise security, dedicated contact and custom integrations |
Eigent states that 10% of monthly subscription revenue is donated to CAMEL-AI.org’s open source work.
Pricing reflects information as of August 2026. Note that the official pricing page shows the same amount as Plus in the Free plan column — an inconsistency you can verify yourself — so we recommend confirming the actual billing terms on the signup screen. For current pricing, check the official site.
Pros and Cons
✅ Pros
- You can switch to your own API keys, making costs predictable relative to your usage
- Open source (Apache 2.0), so you can deploy it on your own infrastructure with no license fees
- Splitting work across multiple agents lets you handle more complex tasks in one pass than a single agent could
- Adoption spread quickly right after release, so you have development momentum and community activity as signals to judge by
⚠️ Cons
- The high-credit plan (Pro) runs $129.99/month ($99.99/mo even billed annually), which is steep for personal use
- Multi-agent setups have their own configuration and prompt design that takes time to learn
- Exposing Eigent itself as an MCP server callable from outside is still only a feature request, not shipped
Comparison with Similar Services
| Comparison | Eigent | Claude Cowork | CrewAI | AutoGen |
|---|---|---|---|---|
| Provider | EIGENT UK LTD | Anthropic | CrewAI | Microsoft |
| Delivery model | Desktop app (OSS) | Fully managed, cloud-only | Framework, code-first | Orchestration library for developers |
| Target user | General users who want everything in a GUI | Users in the Anthropic ecosystem | Developers | Developers |
Who It’s For
- People who want research reports spanning multiple sources generated automatically
- Anyone running a development workflow where coding, documentation, and data analysis need to happen in parallel
- Teams at companies that can’t send internal data outside and need local or self-hosted gateway operation
- People who want routine work automated continuously via scheduled execution
Summary
Eigent is an open source tool that takes CAMEL-AI’s multi-agent framework as its foundation and delivers parallel collaboration between role-specialized agents in desktop app form. Being able to choose local execution or BYOK makes data privacy easier to maintain, and the barrier to entry is lower than framework-style multi-agent tools that require writing code. On the other hand, Pro and above get expensive, and multi-agent configuration takes some getting used to — so the practical move is to start on the Free plan and see whether it fits your workflow.