Question 1

What is Agentuity?

Accepted Answer

Agentuity is a full-stack, agent-native cloud platform for building and deploying AI agents. It provides everything you need—runtime, storage, sandboxes, observability, evals, and an AI gateway—in one integrated platform, so your team can focus on building agents instead of managing infrastructure.

Question 2

How is Agentuity different from running agents on AWS or Azure?

Accepted Answer

Traditional clouds have some or all of the building blocks, but you're wiring them together yourself—IAM roles, security groups, separate billing for each service. Agentuity is purpose-built for agent workloads: long-running, stateful processes that need integrated storage, sandboxes, and observability out of the box. One platform, one SDK, one bill.

Question 3

Do I need to use a specific AI framework?

Accepted Answer

No. Agentuity is framework-agnostic. Bring your own framework—Mastra, AI SDK, LangChain, LangGraph, CrewAI, or plain code. Agentuity provides the infrastructure layer underneath, not the orchestration layer on top.

Question 4

What LLM providers does Agentuity support?

Accepted Answer

Agentuity's AI Gateway supports OpenAI, Anthropic, Google, Groq, Mistral, xAI, and more. You can use Agentuity's managed API keys with unified billing, or bring your own keys. Switch providers without changing your agent code.

Question 5

Is Agentuity free?

Accepted Answer

Agentuity accounts are free—you only pay for usage. Every account starts with $5 in free credits, no credit card required. There are no paid tiers or seat-based pricing.

Question 6

How fast can I get an agent to production?

Accepted Answer

About three minutes. Run one command to scaffold a project, write your agent logic, then deploy it to production. No infrastructure configuration, no Docker files, no CI/CD setup required.

Question 7

Is Agentuity open source?

Accepted Answer

The Agentuity SDK is open source under the Apache 2.0 license. You can audit the code, contribute, and use it freely. The platform is designed for zero lock-in: bring your own LLM, bring your own framework, deploy anywhere.

Question 8

Can I deploy agents on my own infrastructure?

Accepted Answer

Yes. The Gravity Network lets you deploy the same Agentuity platform to your VPC, on-premises infrastructure, or edge devices. Same SDK, same services, same observability—wherever your data needs to stay.

Question 9

How does Agentuity handle security?

Accepted Answer

Sandboxes run in isolated containers with strict resource limits and network disabled by default. They can access your project's production services (KV, Vector, Storage) when needed, but cannot access other projects or sandboxes. All communication uses end-to-end TLS. For sensitive workloads, deploy to your own VPC with the Gravity Network for full data sovereignty.

Question 10

What happens if my agent runs for hours?

Accepted Answer

That's exactly what Agentuity is built for. Unlike traditional serverless which times out, the Agentuity runtime is designed for long-running, stateful workloads. State persists across sessions. You pay for actual compute, not idle time.

Question 11

What languages and runtimes does Agentuity support?

Accepted Answer

Agentuity's SDK is TypeScript-first, running on Bun for maximum performance. Sandboxes support multiple runtimes including Bun, Python, Node.js, and Golang. The platform provides end-to-end type safety from agent schemas to React frontend hooks.

Question 12

How does the AI Gateway work?

Accepted Answer

The AI Gateway routes LLM calls to any supported provider without requiring API keys in your code. It provides unified billing across all providers, higher rate limits, and detailed telemetry including token usage, costs, latency, and span-level debugging via OpenTelemetry.

Question 13

What storage options are available?

Accepted Answer

Agentuity provides Redis KV for caching and session state, PG Vector for semantic search and RAG, S3-compatible object storage via Tigris for globally distributed files and artifacts, Neon Postgres for relational data, and durable streams for persisting streaming data. All accessible from agent context via ctx.kv, ctx.vector, and ctx.storage without infrastructure setup.

Question 14

How do evals work in production?

Accepted Answer

Unlike most eval tools that only test at build time, Agentuity runs evals on every production session. Evals run after the agent has responded, so they never add latency to user-facing requests. They test the complete input and output of your agent—including multiple LLM calls, tool executions, and orchestration steps. Results appear as spans in your OpenTelemetry traces.

Question 15

What is Workbench?

Accepted Answer

Workbench is a schema-aware testing environment for AI agents. It reads your agent schemas and auto-generates the right input forms. Test local and deployed agents with the same interface—no hand-crafting JSON payloads. Every run emits detailed telemetry with jump-to-trace capability.

Question 16

How do sandboxes work?

Accepted Answer

Sandboxes are isolated execution environments for running untrusted code safely. One function call creates a sandbox, runs your command, and cleans up automatically. Network is disabled by default with strict resource limits. Create snapshots to save filesystem states for faster cold starts.

Question 17

Can I build frontends for my agents?

Accepted Answer

Yes. Deploy React apps with end-to-end type safety flowing from your agent schemas to your hooks. Agent schemas produce typed hooks via @agentuity/react. Streaming responses, WebSocket connections, and SSE are pre-configured. Deploy alongside your agents or separately on any hosting platform you prefer.

Question 18

How does observability work?

Accepted Answer

Agentuity provides automatic OpenTelemetry instrumentation for all LLM calls, API calls, and storage operations. See waterfall spans grouped by session, with timing, prompts, responses, and costs. Logs and eval results attach to the same timeline—no separate tools needed.

Question 19

What are custom domains?

Accepted Answer

Point your domain at any agent or frontend. SSL certificates, DNS routing, and load balancing are handled automatically. Define custom domains in your project config—version controlled and deployed atomically with your agents. Full global edge network with DDoS protection.

Question 20

How does agent-to-agent communication work?

Accepted Answer

Call other agents with one line of code regardless of framework. Agents can communicate across any framework or language—a Mastra agent can call a LangGraph agent seamlessly. Agentuity handles routing, serialization, and type safety between agents.

Frequently Asked Questions

General

Technical

Frequently Asked Questions

General

Technical