nesquena/hermes-webui

Features

• Autonomous Agent Orchestration - Provides a platform for orchestrating autonomous agents that execute shell commands and manage persistent memory.
• On-Premise Deployment - Runs the entire AI orchestrator and its data on local hardware to ensure total auditability and data sovereignty.
• Agent Management Interfaces - Provides a comprehensive browser-based graphical interface for configuring, monitoring, and interacting with autonomous AI agents.
• Autonomous Task Scheduling - Implements a built-in cron system for scheduling and managing recurring autonomous agent background tasks.
• Provider-Agnostic Adapters - Provides a provider-agnostic adapter layer to switch seamlessly between cloud APIs and local LLM instances.
• Model Context Protocol - Utilizes a standardized protocol to connect AI models with local data sources and external tools for improved context.
• External Service Integrations - Enables agents to connect to and utilize third-party services and tools through agent-driven tool integration.
• MCP Protocol Integrations - Implements the Model Context Protocol to integrate external tools and capabilities into AI agents.
• Global Agent Memory - Stores learned facts and conventions indefinitely across all projects without repository boundaries.
• Interchangeable Model Support - Deno X uses various local or cloud-based models from different providers interchangeably.
• LLM Gateways - Acts as a centralized gateway for aggregating and swapping multiple LLM providers via a single interface.
• LLM Provider Adapters - Features standardized interfaces for connecting and swapping between multiple cloud and local LLM providers.
• Headless Execution Pipelines - Provides a headless execution mode for running AI agent tasks on a server without a GUI.
• Markdown Memory Stores - Stores user profiles and conversation history in human-readable markdown files for transparency and portability.
• Knowledge Graph Builders - Builds a structured knowledge graph of facts that accumulates across different projects and sessions for long-term recall.
• Agent Memory Management - Uses markdown files and knowledge graphs to provide a persistent memory system for long-term agent context.
• Workspace Action Execution - Implements a workspace action execution framework to run code, edit files, and call external APIs.
• Cron Scheduling - Includes a built-in cron system for executing autonomous background tasks on a fixed timetable.
• Shell Command Execution - Enables AI agents to interact with private databases and local files using the host server shell.
• Sandboxed Shell Executions - Requires explicit user authorization before executing potentially dangerous shell commands on the host.
• Private Infrastructure Hosting - Provides infrastructure for deploying AI agents within private networks to ensure full data sovereignty.
• Self-Hosted AI Infrastructure - Enables the deployment of LLM interfaces and orchestrators on private hardware for data sovereignty.
• Self-Hosted AI Platforms - Serves as a self-hosted AI platform for executing background jobs and managing agent memory on private hardware.
• Self-Hosted Services - Enables autonomous task scheduling on self-managed private infrastructure to ensure data sovereignty.
• Agent Memory Synchronization - Synchronizes a single agent and its persistent memory across various chat apps and social networks.
• Daemonized Process Persistence - Operates as a persistent daemon process to ensure background tasks and interfaces remain always-on.
• Shared Knowledge Graph Memory - Utilizes a structured knowledge graph to maintain long-term factual memory across different projects and sessions.
• Background Task Scheduling - Executes scheduled autonomous jobs as background processes on private local hardware.
• Web Chat Interfaces - Offers a web-based chat interface for organizing conversational sessions, projects, and tags.
• Active Process Redirection - Allows users to inject course-correction messages into a running autonomous process without interrupting its progress.
• External Memory Integrations - Connects to external memory backends to enable persistent long-term recall and structured fact extraction.
• Hierarchical Task Delegation - Implements hierarchical task delegation by spawning parallel sub-agents to handle specialized workloads.
• Autonomous Capability Evolution - Enables agents to update their own functional capabilities automatically over time without manual plugin installation.
• Agent Configuration Profiles - Supports managing agent configuration profiles to switch between custom endpoints, API keys, and skill sets.
• Agent Persona Definitions - Implements configurations for defining behavioral traits, system prompts, and operational identities for AI agents.
• Model Provider Integrations - Provides unified interfaces for connecting to custom cloud endpoints or local model instances.
• Agentic Task Orchestrators - Orchestrates specialized sub-agents to decompose complex objectives into smaller, manageable tasks.
• Agentic Tool Orchestration - Manages the discovery and execution of external third-party tools to incorporate results into agent memory.
• Action Approval Gates - Implements security gates that require explicit user approval before executing potentially dangerous shell commands.
• Model Request Routing - Directs API requests to various AI backends, including both cloud-based and local deployments.
• Language Model Integrations - Uses adapters and streaming interfaces to connect and swap between various proprietary or local language models.
• Hybrid Provider Integrations - Integrates with AI model providers via API keys for both cloud and local server environments.
• Context Compression - Summarizes long conversation histories to keep the context within the model's token limits.
• Chat-to-CLI Bridging - Bridges the gap between CLI sessions and the web interface by importing conversation history.
• External Tool Integrations - Connects AI assistants to external communication tools like Slack and Telegram.
• Model API Providers - Provides unified API access to multiple large language model providers using keys or OAuth.
• Hot-Swappable Providers - Allows switching between different AI model providers at runtime via configuration without losing state.
• Model Context Protocol Implementations - Implements the Model Context Protocol to extend agent capabilities with standardized external tools.
• Execution Environment Orchestration - Executes agent tasks across diverse environments including local shells, Docker containers, and SSH connections.
• Self-Improving Automation Code - Implements the autonomous generation and sharing of new helper functions to resolve missing agent capabilities.
• Web Search Tools - Provides integrated web search tools for agents to retrieve real-time information and synthesize results for recurring tasks.
• User Preference Management - Persists user preferences and conversation history across sessions to enable context-aware interactions.
• Real-Time Text Streaming - Implements real-time text streaming using server-sent events to display LLM tokens incrementally as they are generated.
• External Process Delegation - Delegates specific coding tasks to external CLI tools and captures the output back into the agent workflow.
• Multi-Channel Support Inboxes - Supports interaction across multiple communication channels including WhatsApp, Telegram, and Discord.
• Interval-Based Task Execution - Triggers automated agent tasks at precise, fixed time intervals on a self-hosted server.
• Daemonized Session Persistence - Operates as an always-on server process to maintain agent state and availability without an active user session.
• Agent Function Libraries - Ships a library of pre-built functions for web search, browser automation, and vision analysis.
• Slash Command Interfaces - Features a slash command system with an autocomplete menu to trigger internal functions and system tasks.
• Execution Sandboxes - Executes agent processes within isolated Docker, SSH, or serverless sandboxes for secure code execution.
• Recurring Job Scheduling - Automates recurring jobs on a fixed timetable using CRON expressions on self-hosted hardware.
• Self-Hosted Infrastructure - Supports deployment on self-managed infrastructure to maintain full control over data and agent operations.
• Chat Organization - Organizes AI conversations into named projects and tagged groups for efficient retrieval.
• Chat Platform Integrations - Synchronizes a single agent and its memory across web interfaces and messaging platforms like Telegram and Discord.
• Messaging App Integrations - Enables sending and receiving messages across diverse platforms including Discord and Telegram.
• Token Streaming - Streams LLM tokens incrementally from the backend to the browser for real-time response rendering.
• Private Hosting - Ensures conversation history and agent data remain under private control via local hardware hosting.
• Agent Process Isolation - Isolates agent execution environments using container boundaries and approval workflows to protect the host.
• Dynamic Skill Synthesis - Writes new skill modules to extend functional abilities based on learned context and operational experience.
• Conversational Session Managers - Implements a conversational session manager that allows users to reopen the most recent conversation with full history and context.
• Cross-Platform Desktop Applications - Includes a cross-platform desktop application to synchronize configurations, sessions, and agent memory.

Star history