QA That Proves It
QA that reads your code, drives your product, and proves what broke, with the request and the screenshot attached. Plus docs and projects that stay true to the code.
Artel has two surfaces over one body of knowledge. Artel QA validates the running product; Artel Work is where the documentation, decisions and projects live. Both read from the same graph, which is why a test finding can name the spec it contradicts, and a document can be flagged stale by the code that moved underneath it.
A black-box tool can tell you a button exists. It cannot tell you the button is supposed to write a row to a particular table, that the write travels through a specific chain of services, or that the row it wrote instead was the wrong one. Artel QA can, because it read the code first, then drove the product like an engineer would, and kept the evidence.
Artel binds directly to GitHub repositories. Commits and issues are mirrored into the workspace in real time, and individual commits can be linked to specific docs and tasks. Code context flows into the AI assistant automatically, so it understands what’s been built, and not only what’s been written down.
Every document, task, and linked commit becomes part of the workspace the assistant can draw on. Ask a question and the Pro assistant answers with citations from the project itself, no invented context, no bleed between workspaces. As the project grows, so does what the assistant knows.
Artel connects natively to Flow Console, so AI agents can do real work in your workspace: create documents, update tasks, log decisions, and query project state, with each action permission-scoped and attributable. Your workspace becomes something agents can act on, instead of only read.
Artel is live and in active use. All core surfaces, documents, tasks, whiteboards, GitHub integration, and the Pro AI assistant, are shipped and available today.
Artel doesn't ship in isolation. It plugs into the rest of the Metantel line-up through one shared protocol surface, context flows in and out without leaving the family.
Runtime evidence from a QA run flows back to the coding agents, so they act on what actually failed rather than what static analysis suspects.
Open Flow Console →Mature Foundry conversations spin directly into Artel workspaces with context and collaborators intact.
Open Foundry →