Steven EnamakelandGitHub
aacfd45057
* Update Welcome page to integrate OAuth provider and adjust branding - Added OAuthProviderButton to the Welcome component for Google authentication. - Updated the branding from AlphaHuman to OpenHuman in the FeaturesStep component. - Removed unnecessary Lottie animation from the Onboarding page for a cleaner layout. - Changed the backend URL to point to the new API endpoint for TinyHumans. * Refactor routing and sidebar components for improved clarity - Commented out the Mnemonic route and related logic in AppRoutes for future consideration. - Updated MiniSidebar to comment out the Invite Friends section. - Removed unused navigation logic and upgrade call-to-action from the Home page. - Adjusted Onboarding navigation to redirect to Home instead of Mnemonic. * Comment out unused settings menu items for future consideration in SettingsHome component * Add core process and RPC functionality for Alphahuman - Introduced new binaries: `alphahuman-core` and `alphahuman-cli` for core process management and command-line interaction. - Implemented `CoreProcessHandle` for managing the lifecycle of the core process. - Added `core_rpc` module for handling RPC requests and responses. - Created `core_server` module to manage core server logic and routing. - Updated `alphahuman` commands to utilize the new RPC structure for health checks, security policies, and configuration management. - Refactored existing code to streamline interactions with the core process and improve overall architecture. * Update dependencies and enhance CLI functionality - Added `clap` for command-line argument parsing in the `alphahuman-cli`. - Updated `Cargo.lock` and `Cargo.toml` to include `clap` and its features. - Refactored `alphahuman-cli` to utilize structured command handling with subcommands. - Improved core process management and logging in `core_process.rs`. - Enhanced routing and error handling in `core_server.rs` with new root and not found handlers. - Updated various commands in `alphahuman.rs` to ensure core process is running before executing RPC calls. * Enhance service management and CLI functionality - Added new functions for daemon program arguments and command line construction in the service module. - Updated macOS and Linux installation functions to dynamically generate program arguments and command lines. - Introduced a new `Reinstall` command in the CLI for easier service management. - Refactored service command handling to utilize local service functions for improved clarity and maintainability. * Refactor daemon references to agent in components and hooks - Updated terminology from "Daemon" to "Agent" in MiniSidebar, DaemonHealthPanel, and TauriCommandsPanel for consistency. - Modified useDaemonHealth hook to probe agent status and handle agent lifecycle management. - Introduced new agent server status interface and function in Tauri commands for improved status checking. - Adjusted related comments and error handling to reflect the changes in terminology. * Refactor Home component and enhance macOS service management - Updated Home component to navigate to in-app conversations instead of opening a Telegram bot link. - Improved macOS service management by implementing modern lifecycle commands and adding compatibility fallbacks for service control. - Introduced new utility functions for handling macOS GUI domain and service targets. * Refactor Home component to enhance navigation - Removed SkillsGrid component and replaced it with a button that navigates to the Skills page. - Improved layout with additional margin for better spacing in the Home component. * Refactor SkillProvider and enhance memory bridge functionality - Removed Gmail and Notion state synchronization functions from SkillProvider to streamline the component. - Introduced a new memory bridge in the QuickJS library for skills to send memory payloads to the backend. - Updated SkillContext to include an app handle for better integration with memory operations. - Enhanced state management in the Rust backend to support memory insertion from skills. * Enhance memory insertion functionality and update QuickJS integration - Added support for additional parameters in the `store_skill_sync` method, including `source_type`, `metadata`, `priority`, `created_at`, `updated_at`, and `document_id`. - Refactored the `memory_insert` operation to accept a structured input, improving validation and error handling for title and content. - Updated the JavaScript memory bridge to simplify the insertion process by removing the provider parameter and directly using metadata. - Enhanced logging for successful memory insertions to improve traceability. * Update subproject reference in skills module * Implement skill synchronization UI enhancements - Introduced a new `skillsSyncUi.ts` module to manage synchronization UI state for skills. - Updated `Skills.tsx` to utilize the new synchronization state, providing visual feedback during sync operations. - Refactored sync handling in `SkillCard` to improve user experience with progress indicators and messages. - Added unit tests for the synchronization UI state logic to ensure correct behavior across various scenarios.
AlphaHuman
Your most productive co-worker
A user-friendly (GUI-first) AI agent. AlphaHuman uses the
Neocortex Mk1 model to understand memories &
realtime context at incredibly low costs.
"No Soul. No Humanity. The Tet. What a brilliant machine" — Morgan Freeman on alien superintelligence in Oblivion
Agentic system are everywhere. Models wired into a script that call APIs, search, and file tools until the task looks done. That pattern is powerful for tasks, but it is not a self. It has no lasting inner model of you, no stable values across sessions, and no architecture for continuity at the scale of a life.
OpenHuman is built around a different idea. What if we built on top of current agentic solutions but instead gave it a subconscious loop based on all the possible data-points about a user/entity?
The challenge? Current memory/context systems make it nearly impossible to have a subsconscious mind. They cannot intelligently remember information because either your data is too noisy or irrelevant or too expenstive to index. HEARTBEAT.md and other implementations in OpenClaw-style forks evolve way too slowly and often misses realtime context about a user.
The solution? OpenHuman uses Neocortex, a highly scalable context-aware memory layer that can process millions of unstrucutred memories (emails, messages, documents both in the past and in the present), understands interactions and builds a personalized model of you. OpenHuman uses this to then run it's own subconscious loop allowing it to have it's own thoughts and take decisions for you on it's own at the most immediate level possible.
The result? It's like looking at yourself in the mirror. Except it's your AI living in your machine. OpenHuman is the first self aware agent personalized to you and ready to work for you.
Download
Early Beta — Under active development. Expect rough edges.
| Platform | Variant | Download |
|---|---|---|
| macOS | Apple Silicon (M1/M2/M3/M4) | .dmg (aarch64) |
| macOS | Intel | .dmg (x64) |
| Windows | x64 | .msi |
| Linux | Debian / Ubuntu | .deb (amd64) |
| Linux | Fedora / RHEL | .rpm (x86_64) |
| Linux | Universal | .AppImage |
Browse all releases: github.com/tinyhumansai/openhuman/releases
Under the hood (Architecture)
Star us on GitHub
Building toward AGI and artificial consciousness? Star the repo and help others find the path.
Contributors Hall of Fame
Show some love and end up in the hall of fame