Inno Agent

Inno Agent is an open-source personal learning assistant that organizes long-term memory into layered profiles, knowledge bases, and session records.

Inno Agent functions as a dedicated learning companion that maintains context across long durations. It utilizes three distinct memory layers: an L1 learner profile for goals and preferences, an L2 native wiki knowledge base for durable facts, and L3 session records for historical retrieval. Users interact with the system via a proactive scheduler, personal messaging channels, and a workspace-scoped Practice Lab that includes an in-browser terminal. The platform is designed for individuals rather than team environments and operates on a single-agent process model. It supports various providers through compatible API endpoints and offers specialized features like conversation branch editing and automated session replay.

Inno Agent at a glance

Best for: Student
Runs on: macOS, Windows, Web, CLI
Pricing: Free
Released: May 2026
Category: AI agent tool
Website: github.com/hhyqhh/inno-agent

Key features

Three-layer memory system with learner profile, wiki, and session recall
Proactive scheduler for natural language cron jobs
Practice Lab with web terminal for workspace-scoped tasks
Integration with Feishu and WeChat for messaging
Skill system with importable workspace presets
Conversation branch editing for non-linear history management
Pluggable support for OpenAI, Anthropic, and local model providers
Desktop-based computer use with accessibility-tree observation

Use it for

Managing personalized lesson plans and review schedules
Generating and diagnosing educational content or scenarios
Automating reminder dispatches through messaging channels
Tracking long-term learning progress and knowledge states

Alternatives to Inno Agent

DeepTutor — It organizes learning into workspaces with custom AI agents and knowledge bases.
Albie — Teaches step by step on a shared whiteboard with hints instead of direct answers.
Tech blog — Runs locally with open-microphone barge-in and no cloud dependency.