Deploy a self-improving AI worker. Neon learns skills from past tasks, operates autonomously via browser control, and keeps costs protected via strict budget governance.
Every command is analyzed for safety and cost before execution. Neon monitors API expenses dynamically to prevent unexpected costs.
Neon learns from successes and saves them as reusable "Skills" in its procedural database. Future similar tasks execute 10x faster.
Integrated Playwright browser control allows Neon to autonomously research web links, fill forms, and download documentation.
What is Neon? Neon is a next-generation autonomous AI workspace agent. It features combined architectures of Management (Paperclip), Learning (Hermes), and Adversarial (OpenClaw) frameworks.
Budget Protection: Neon guarantees budget compliance. You set strict limits on monthly LLM consumption, and Neon will pause operations before exceeding your threshold.
Zero-Config Web Interface: Manage active models (Gemini, Groq, OpenRouter), toggle tools, review costs, and manage system prompts from a local, single-click web dashboard.
Full ownership of the custom agent codebase, Docker setup, and local dashboard UI. Includes setup call, custom skills library integration, and 1 year of security support.
Neon combines three powerful frameworks: governance/budget control from Paperclip, self-improving learning loops from Agent Hermes, and adversarial multi-agent quality control from OpenClaw - all in one agent.
When Neon successfully completes a task, it analyzes the approach and saves it as a reusable 'Skill' in its procedural memory. Future similar tasks automatically match and use these learned skills.
Yes, Neon features a zero-config web interface. Click the SETTINGS button in the dashboard to configure API keys, switch LLM models, and adjust parameters.
Neon supports gemini/gemini-1.5-flash, groq/llama-3.1-70b, openrouter/deepseek/chat, and more - dynamically switchable.
After every execution, a Critic-Reviewer loop evaluates the output using the SHARP framework (Specific, Honest, Actionable, Relevant, Precise) and refines if needed.
Yes, Neon comes with a Docker configuration for one-click VPS deployment. It serves the Web UI on port 5000 and connects to Telegram/Discord bots automatically.
The governance layer analyzes every user command for safety risks and estimated API costs before execution. If a command exceeds budget or triggers safety concerns, it's rejected.