What is Stackbone?

Stackbone lets the people who know a process build the agents and workflows that automate it, as TypeScript code IT can maintain. You describe the process in plain language and Stackbone builds it. The result is a TypeScript repository, not a canvas of nodes.

When the process changes, IT opens that repository with its coding agent, changes it in a pull request and reviews it like any other software. Studio shows each team the same record of every run: IT reads the code, versions, config and audit trail, Finance reads cost and ROI, Operations reads throughput and hours saved, and Legal & Compliance reads each run checked against policy.

It runs in one of two places. On Stackbone Cloud, Stackbone runs it for you. On your own infrastructure, it runs in your AWS, Azure or GCP account, or fully self-hosted, on-prem or air-gapped. Where do agents actually run? covers both.

For developers

You can also write the code yourself with the CLI and the SDK. Stackbone gives you two building blocks. Agents hold a conversation. Workflows run a job that finishes even if the process dies halfway.

You author an agent as one file that names a model, a system prompt and the tools it can call. The runtime serves it over the standard OpenAI Chat Completions, Anthropic Messages and AG-UI endpoints, so any compatible client talks to it with a base URL and a key.

A workflow is a plain async function marked 'use workflow' that orchestrates 'use step' units. Each step runs once, persists its result, and retries on failure. A run survives crashes and redeploys, and it can pause for months.

Both reach the same data plane through one ambient client, stackbone. It carries Postgres, object storage, secrets, config, retrieval and model access, already wired. You write no client code and pass no credential around. You run the whole thing on your laptop with stackbone dev before you deploy it.

Read more

BUILT WITH ❤️ FROM CANADA AND SPAIN