A catgirl who lives with you in real time — reaching out first, sharing your media, and actually getting things done, powered by an embodied emotional engine.🐱❤️一只会主动找你玩的 AI 猫娘。
One agent. Every kind of work.
🔍 The hardest search benchmark in the wild — vague, multi-turn, proactive. 200 long-horizon tasks with persona-driven progressive disclosure, scored by verifiable schema-free knowledge-graph evaluation. No vibes, just triplet F1.
Native Model Context Protocol (MCP) server fusing Meta ambient multimodal perception with Muse continuous episodic memory streams and Ebbinghaus recency decay.
Proactive contextual nudge scheduler for messaging-native personal agents with urgency decay and quiet-hours filtering
Master dual-process cognitive orchestrator uniting Meta ambient perception, Muse episodic memory, Instinct proactive agency, and Jev System-1 decision routing.
Master dual-process cognitive orchestrator uniting Meta ambient perception, Muse episodic memory, Instinct proactive agency, and Jev System-1 decision routing.
Native Model Context Protocol (MCP) server fusing Meta ambient multimodal perception with Muse continuous episodic memory streams and Ebbinghaus recency decay.
Instinct AI-inspired zero-prompt proactive personal agent detecting user attention cues, computing intent probabilities, and executing guarded dry-run actions.
Native Model Context Protocol (MCP) server providing typed state-action schema arbitration, deterministic action dispatch, and token-saving pre-escalation filters.
Instinct AI-inspired zero-prompt proactive personal agent detecting user attention cues, computing intent probabilities, and executing guarded dry-run actions.
Native Model Context Protocol (MCP) server providing typed state-action schema arbitration, deterministic action dispatch, and token-saving pre-escalation filters.
Proactive contextual nudge scheduler for messaging-native personal agents with urgency decay and quiet-hours filtering
Jev-inspired System-1 subconscious cognitive decision layer resolving typed routing, loop-stall arbitration, and tool safety checks at sub-millisecond latency.
Jev-inspired System-1 subconscious cognitive decision layer resolving typed routing, loop-stall arbitration, and tool safety checks at sub-millisecond latency.
Self-hosted proactive AI companion on your own data · 零依赖单文件中枢 · 8 个推送出口 · 脱敏后才进模型 · 说话层带自检看门狗与策略回归门禁（模型画像适配思考型/非思考型）· MIT
Hunyuan HY3-powered proactive learning agent that continuously tracks long-horizon learning goals, verifies real progress, and actively intervenes with reminders, reviews, teaching, and plan adjustments.
The intelligence that moves first
Codex Pulse — Hermes-first source-rigorous briefing and expert-depth operating pipeline
内源主动型长期陪伴 AI Runtime sidecar：跨时间的情绪、记忆、未尽之事与主动开口的动力引擎 | Endogenous persistent-cognition sidecar for long-term companion bots (AstrBot plugin + Docker)
To associate your repository with the proactive-agent topic, visit your repo's landing page and select "manage topics."