Website | Documentation | Issues | License: MIT
A voice-first AI assistant that sees through your camera, controls your browser and desktop, remembers everything, learns from experience, and runs tasks on a schedule. No build step. No framework. Open source.
curl -fsSL https://appatalks.github.io/eva-agent/get-eva.sh | bashOr clone and run locally:
git clone https://github.com/appatalks/eva-agent.git
cd eva-agent
./install.sh # install dependencies
cd standalone && npm install && npm run dist
./dist/'Eva Standalone-5.2.3.AppImage'Prereqs: Node.js 24+, Python 3.12+, GitHub Copilot CLI (copilot auth login).
| Camera vision | Webcam presence sensing, face-detection auto-wake, on-demand "look" with gpt-4o |
| Browser agent | Playwright-based DOM control, persistent Chrome login, hybrid vision fallback |
| Desktop agent | PyAutoGUI mouse/keyboard control, optional AT-SPI via computer-use-linux MCP |
| Voice interface | Full-screen voice orb, wake/barge-in, TTS (OpenAI, Polly, Bark, browser) |
| Signal messaging | Send-only text notifications via signal-cli, keyword-triggered or on-demand |
| Persistent memory | Kusto/ADX or local SQLite: conversations, emotion tracking, semantic recall |
| Self-improving skills | Auto-extracts reusable skills from successful tasks, stored as drafts |
| Cron scheduler | Standard cron expressions, recurring prompts, morning briefings, alerts |
| Subagent parallelism | Spawn up to 4 concurrent ACP tasks, results via notifications |
| Multi-provider | OpenAI, Google Gemini, GitHub Copilot, lm-studio (local) |
| Doctor diagnostics | Structured readiness probe for every subsystem with actionable fixes |
| MCP ecosystem | Azure, GitHub, Kusto, computer-use-linux desktop control |
| Cognitive layer | Eva + Reviewer dual-agent pipeline with configurable models |
| Dual data mode | Cloud (Copilot CLI + MCP) or Local (LM Studio + direct MCP, fully offline) |
Select Eva (AIG) in the model dropdown for the full experience.
For persistent memory, point Settings > MCP at an Azure Data Explorer cluster, or use the default local SQLite backend (zero setup). For semantic recall, add an OpenAI key in Settings > Auth (falls back to keyword matching without one).
For Signal notifications, install signal-cli and link it to your Signal account (signal-cli link -n "Eva"). Enter sender and recipient numbers in Settings > Auth.
Import skills from text, URLs, GitHub repos, or files in Settings. Eva normalizes them into her format, stores in ADX, and applies matching skills automatically.
- README-2.md: architecture, MCP, ACP, browser-only setup, roadmap
- standalone/README.md: AppImage build and runtime
- Website: features, comparison, install guide