- CollectionAgent Runtime & Infrastructure
- Classificationagent-native
- MCPUnavailable
- Agent SkillUnavailable
- Interfaceswebhook · CLI · SDK · http · API
What it does
Agentuity is a deployment and runtime platform for AI agents as workloads: type-safe APIs and optional React frontends, Postgres / Redis / vector / object storage exposed to agents as tools, ephemeral sandboxes for untrusted code, OpenTelemetry-based observability (prompts, responses, cost per span), evals tied to production traffic, and I/O integrations (email, SMS, webhooks, cron). It targets long-running, session-oriented agent execution rather than classic stateless HTTP services — see AI agent infrastructure and related architecture guides on the marketing site.
Agent-facing interfaces
- Catalog protocols: webhook, CLI, SDK, http, API
- MCP: ⚠️ No dedicated MCP server documented as a first-class product surface.
- Agent Skill: ⚠️ No official SKILL.md registry entry located yet.
- Onboarding type:
sdk
How agents onboard
# Follow quickstart at https://agentcompany.com/agentuity — install CLI and SDK, then deploy.
# Example pattern (see official docs for current commands):
npm install @agentuity/sdk
See the full onboarding section in the dossier.
Protocol surface
| Interface | Detail |
|---|---|
| TypeScript SDK | @agentuity/sdk — see GitHub |
| Python SDK | agentuity — see GitHub |
| CLI | agentuity CLI — see GitHub |
| HTTP / Webhooks | Product I/O and API routes — see https://agentcompany.com/agentuity |