Built for autonomous AI
Everything you need to run intelligent agents on your own infrastructure. No compromises.
Chat
Conversational AI interface with streaming responses, tool execution, and multi-turn context. Chat with any agent using your preferred LLM provider.
- Streaming responses with real-time token generation
- Inline tool execution — agents act while you chat
- Multi-turn conversation with full context retention
- File attachments, image uploads, and rich media support
- Per-conversation agent and model selection
Snippbots & Orchestration
Create distinct AI personas with custom system prompts, model selection, and autonomous tool execution. Run multiple agents in concert — sub-agents handle specialized tasks while a supervisor coordinates.
- Distinct personas with custom names, system prompts, and behavior profiles
- Per-agent model selection across 9 LLM providers
- Sub-agent spawning with specialized roles (researcher, coder, reviewer, and more)
- Inter-agent messaging and result aggregation (concatenate, merge, pipeline, AI summarize)
- VM-based isolation with Docker, Podman, or process sandboxing
- Shared and isolated memory scopes per agent
Projects
Organize agent work into projects with task planning, phase management, and execution tracking. Generate plans, assign agents, and monitor progress through a visual DAG.
- Project creation with goals, phases, and task breakdown
- Plan generation and iterative refinement
- Task assignment to specific agents
- Visual DAG for dependency tracking
- Execution metrics and audit trail
Persistent Memory
Multi-tier cognitive memory that gives agents real memory across sessions — sensory buffer, working memory, episodic store, knowledge graph, and meta-cognitive reflection, with vector and full-text retrieval over the whole stack.
- Episodic memory — stores experiences chronologically with importance scores
- Semantic/vector search — find relevant memories by meaning, not keywords
- FTS5 full-text search — fast keyword-based retrieval
- Knowledge graph — entities, relations, and structured world knowledge
- Intelligent forgetting — gradual decay and importance-based retention
History
Complete execution history for every agent task. Track outcomes, duration, token usage, and full transcripts — searchable and filterable.
- Task execution logging with success/fail/timeout outcomes
- Duration and token usage tracking per task
- Full conversation transcripts and error messages
- Search and filter by agent, project, date, or status
- Conversation archiving and bulk management
Insights
Proactive, heuristic-based analysis of agent activity. Snippbot surfaces patterns, suggestions, and connections across your work — no LLM needed, fully deterministic.
- Deterministic heuristic analyzer — no LLM cost
- Six insight types: research, suggestions, reminders, connections, warnings, achievements
- Configurable delivery modes and insight queue
- Proactivity settings for insight generation priorities
- Feedback collection to refine future insights
Marketplace
Install and publish Singularity packages — tools, workflows, MCP servers, jobs, channels, and hooks. Permission management, checksum verification, and a publisher dashboard.
- Singularity package format for tools, workflows, MCP servers, and more
- One-click install with dependency resolution
- Permission grant management per package
- Checksum verification and package vetting
- Publisher dashboard for package authors
Assets
Community asset library for storing and managing files, images, and artifacts. Checksum verification, MIME type validation, and tag-based organization.
- File storage with checksum integrity verification
- MIME type validation and size tracking
- Tag-based organization and search
- Asset streaming API for programmatic access
- Browse and manage uploaded files and artifacts
Workflow Builder
Visual DAG editor for building complex automation pipelines. Drag-and-drop step creation with 7 step types, version control, and parallel execution.
- Visual DAG editor in the Aurora UI
- 7 step types: LLM, tool, conditional, loop, approval, sub-workflow, parallel
- Version history and rollback
- Conditional branching and parallel execution
- Approval gates for human-in-the-loop workflows
Job Scheduler
Schedule agent tasks with cron expressions, intervals, or one-time triggers. Full execution history with run logs, retry logic, and job chaining.
- Cron expressions, fixed intervals, and one-time scheduling
- Automatic retry with configurable strategy and max attempts
- Conditional execution — run only when conditions are met
- Job chains — trigger follow-up jobs on completion
- Execution history with output logs and status tracking
Skills
Extend agent capabilities with custom tools and MCP (Model Context Protocol) servers. Create, import, and manage skills with trust levels and encrypted credentials.
- Custom tool creation with full CRUD management
- MCP server integration with trust levels (sandboxed/external/trusted)
- Built-in tool enable/disable preferences per agent
- Encrypted credential storage for authenticated tools
- Import skills from the Marketplace
Hooks
Event-driven automation that triggers actions when things happen. Python scripts, HTTP webhooks, or built-in handlers — scoped to global, project, or agent level.
- Three hook types: Python script, HTTP webhook, built-in handler
- Scope levels: global, project-specific, or agent-specific
- Event filtering — trigger only on matching conditions
- Execution tracking with status, duration, and error logs
- Analytics dashboard for hook performance
Notifications
Webhook-based event delivery built on the hooks system. Wire HTTP endpoints to agent events with automatic retry, HMAC signing, and dead-letter handling.
- HTTP-type hooks with event filtering
- Automatic retry with exponential backoff
- Delivery tracking: pending, processing, success, failed, dead-letter
- Per-hook secret HMAC signing for verification
- Dead-letter queue for failed deliveries
Browser Automation
CDP-powered browser automation with recording, replay, and stealth mode. Agents can browse the web, fill forms, take screenshots, and manage authentication.
- Chrome DevTools Protocol (CDP) integration
- Action recording and replay for repeatable workflows
- Screenshot capture and visual verification
- Stealth mode to avoid bot detection
- Auth manager for credential storage and session handling
Messaging Channels
Connect agents to 6 messaging platforms. Receive commands, send alerts, and have conversations — wherever your team already communicates.
- Slack — slash commands, threads, file sharing
- Discord — bot integration with guild/channel support
- Telegram — bot API with inline keyboards
- WhatsApp — Business API integration
- Microsoft Teams — adaptive cards and bot framework
- Google Chat — Workspace integration with cards
Device Fleet
Pair remote devices for distributed agent tool routing. TOTP-based authentication, device registry, and trusted device management.
- Remote device pairing with TOTP authentication
- Device registry with trust and revocation controls
- Health monitoring with heartbeat and status checks
- Remote tool routing to paired devices
- Token-based session security with rotating refresh tokens
Sandbox Execution
Run agent code safely in isolated environments. Docker, Podman, and process-level sandboxing with GPU support, snapshot/restore, and SSRF protection.
- Docker and Podman container isolation
- Process-level sandboxing for lightweight tasks
- GPU passthrough for ML/AI workloads
- Snapshot and restore for reproducible environments
- SSRF guard to prevent unauthorized network access
Monitor
System health dashboard with token usage tracking, performance history, and security overview. Keep an eye on your agent fleet from a single pane of glass.
- Token usage history and cost tracking per model
- Performance metrics history and trend analysis
- Security overview with audit log
- Connected device and service status
- Daemon uptime and activity log
AI Providers Agnostic
Use your preferred AI model — or your existing Claude subscription. Snippbot is one of the only agent platforms that lets you run agents on your Claude Pro or Team plan with zero additional API costs. 9 providers supported out of the box.
- Claude subscription support — use your Pro/Team plan for agent activity at no extra cost
- Anthropic (Claude) — primary integration with Opus, Sonnet, and Haiku
- OpenAI (GPT-5, GPT-4o, o3, o4-mini)
- Google (Gemini), DeepSeek, Grok, Groq, Mistral
- OpenRouter — unified multi-provider gateway
- Local models via compatible APIs
CLI & API
Full control from the terminal or programmatically. 17 CLI command groups covering daemon, projects, devices, secrets, auth, marketplace, security, and more — plus 800+ REST API endpoints.
- 17 CLI groups: start, stop, status, config, project, channel, device, setup, doctor, reset, secrets, auth, export, security, skills, marketplace, completions
- 800+ REST API endpoints across 16 mounted route groups
- WebSocket support for real-time streaming
- SSE (Server-Sent Events) for live updates
- API key authentication with revocation and audit logging
Beautiful Themes
Gorgeous dark-theme React UI with glassmorphism design and customizable appearance. 27+ pages covering chat, projects, agents, memory, workflows, scheduling, devices, and settings.
- React 18 + TypeScript + Tailwind CSS
- Aurora dark theme with cyan/purple accent palette
- Glassmorphism cards and components
- Customizable interface and appearance settings
- Real-time updates via WebSocket and React Query
Why Snippbot?
The complete self-hosted AI agent platform.
| Self-hosted | ✓ | All data stays on your machine |
| Open source | ✓ | Full source code available |
| Free (personal) | ✓ | No cost for personal use |
| Multi-provider LLMs | ✓ | 9 providers supported |
| Persistent memory | ✓ | 5-tier cognitive architecture |
| Knowledge graph | ✓ | Entities and relations |
| Task history | ✓ | Full execution audit trail |
| Insights | ✓ | Heuristic-based analytics |
| Marketplace | ✓ | Install & publish packages |
| Visual workflows | ✓ | DAG editor with 7 step types |
| Job scheduler | ✓ | Cron + natural language |
| Skills & MCP | ✓ | Custom tools + MCP servers |
| Event hooks | ✓ | Python, HTTP, built-in triggers |
| Notifications | ✓ | Webhook delivery with retry |
| Browser automation | ✓ | CDP with stealth mode |
| 6 messaging channels | ✓ | Slack, Discord, Telegram, and more |
| Device fleet | ✓ | Remote device management |
| Sandbox execution | ✓ | Docker/Podman isolation |
| System monitor | ✓ | Health, tokens, performance |