Msty logo

Msty

Free tier

Private AI for real work

Free tier available·All audiences·Powered by Model-agnostic (local and online models, Azure, Bedrock, Ollama, Llama.cpp, MLX supported)·API available

Key strengths

Privacy-first AI workspaceModel-agnostic (local and online models)Bounded AI agents with reviewable executionGoverned knowledge layer (Knowledge Stacks)Customer-controlled storage and zero product telemetrySSO, audit logs, and role-based access control for teamsSupports local inference engines (Ollama, Llama.cpp, MLX)
Free tier + paid plans · from $149 USD/mo
United States
Self-hostable
No ratings yet

Msty's technical architecture centers on model-agnosticism and customer-controlled infrastructure. Key integration points include:

  • Local inference engines: Ollama, Llama.cpp, MLX
  • Cloud model providers: Azure, AWS Bedrock
  • API access for programmatic interaction
  • MCP Tools integration for extensible agent tooling
  • SSO support for enterprise identity management
  • Audit logs and RBAC for governance and compliance
  • Knowledge Stacks: a governed knowledge layer for structured, reviewable context injection into AI workflows
  • Bounded AI agents with reviewable execution steps, designed to keep humans in the loop
  • Messaging integrations: Discord, Telegram, WhatsApp
  • Self-hostable deployment for full infrastructure control
  • Zero product telemetry — no usage data sent to Msty

The Go product is currently in free beta. Stack pricing is listed as upcoming. Enterprise and Teams plans require a minimum of 5 seats and are available via custom quote.