How is it different from wiring this myself?

Visual workflow tools give you a canvas of nodes. Wiring it yourself gives you an agent, a queue, a Postgres and a secret store to assemble. Stackbone gives you a TypeScript repository your IT team maintains like any other software.

Against a visual workflow tool

In an n8n-style canvas, someone drags nodes until the process works. A year later the canvas holds hundreds of nodes, and whoever changes it first has to reverse-engineer what each one does.

With Stackbone, the person who knows the process describes it in plain language and Stackbone builds it as TypeScript code in a repository. When the process changes, IT opens that repository with its coding agent, makes the change in a pull request and reviews the diff. Tests, code review and version history work the way they already do for the rest of your software.

Against wiring it yourself

You could assemble an agent, a queue, a Postgres and a secret store yourself. Stackbone changes four things about how that project is built.

The wire is standard. There is no proprietary session protocol. A request's model field picks which agent answers, on the same POST /openai/v1/chat/completions and POST /anthropic/v1/messages every agent in the workspace shares. Any OpenAI or Anthropic client already works.

You choose who runs the box. On Stackbone Cloud, Stackbone runs it. On your own infrastructure, Stackbone provisions nothing: you run the container in your own cloud or on-prem, and Stackbone keeps the record of where it lives: its address, its signing secret, and the image tag it runs.

Durability comes with the runtime. A 'use step' function runs once, persists its result, and retries on failure. Kill the runtime mid-run and it resumes from the last completed step.

The image tag is the agent's version. You deploy a new image to ship a change, and you promote nothing in between.

Read more

BUILT WITH ❤️ FROM CANADA AND SPAIN