Skip to main content

About SOAT

SOAT is open-source infrastructure for production-ready AI agents. It is a single Node.js server, backed by PostgreSQL and pgvector, that supplies the parts an agent needs once it leaves a notebook: durable sessions and conversations, multi-agent orchestration, knowledge retrieval over your own documents, long-lived memory, guardrails, IAM, quotas, and traces.

It is self-hosted software, not a SaaS. There is no hosted control plane and no account to create: you run the server and the database, and SOAT calls whichever model provider you configure — OpenAI, Anthropic, Google, Bedrock, Ollama, or any OpenAI-compatible endpoint. This site, soat.ttoss.dev, serves documentation only; there is no API behind this domain.

What it is for​

SOAT earns its place once state, permissions, retrieval, or evidence are involved. A single stateless completion is better served by calling a provider directly. The agent instructions state the boundary precisely, including the cases where SOAT is the wrong answer.

Everything is reachable four ways — the REST API, the MCP endpoint, the @soat/sdk TypeScript client, and the soat CLI — all generated from the same OpenAPI documents, so an operation that exists in one exists in all of them.

Who maintains it​

SOAT is built and maintained by Terezinha Tech Operations (ttoss), a small team that develops it in the open. Development happens entirely on GitHub: issues, pull requests, and design discussions are public, and releases are published to npm and Docker Hub from CI.

License and trademark​

SOAT is released under the Apache License 2.0. You may run it commercially, modify it, and deploy it privately without asking anyone. The name and logo are covered separately by the trademark policy.

Where to go next​