timeback solutions.
Product · throughline

Requirements that justify themselves.

What it is

throughline is a Git-native requirements tool. Every requirement is one small file with a permanent identity, linked to the intent it serves. A single command validates the whole graph and gates your build — so scope can't quietly avalanche, and nothing ships ungrounded.

The files are the product. They live in your repo, in version control, with no external database of record.

The family
Core

throughline

One small file per requirement, each with a permanent identity and typed links to the intent it serves. A single command validates the whole graph and gates CI — orphaned or circular work fails the build.

Composition

throughline-compose

Assemble one project's requirements from many reusable sources. Pull in an industry standard by reference, pinned to an exact edition, without re-typing a line of it — and drift is detected, never silent.

Catalogues

Standard catalogues

Living, grounded requirement sets for the standards teams answer to — OWASP ASVS, GOV.UK, WCAG, NIST AI RMF, NCSC CAF and more. Consume them by reference; upgrade the edition when you're ready.

Jira

throughline-jira

Thread a Jira Cloud issue straight to the requirement it delivers. An issue-panel app links, snapshots at a pinned edition, and flags drift when the requirement moves underneath it.

In development

Host

throughline-connect

The shared host connectors plug into, holding the credentials so the client apps never do. The tools that live in Jira, Git and beyond all reach the same grounded graph through it.

In development

AI

throughline skill for Claude

An agent brief generated from your live project. An assistant proposes grounded changes that a human ratifies — never ungoverned edits, and the provenance is on the record.

See it in action

One council rent-arrears message, composed from six orthogonal content standards. Hold five fixed and swap one — and watch the same message rendered as a web page, a letter and an email, then pitched to a tenant, an officer and a lawyer.

Open the worked examples →

How it fits together

The core proves grounding. Compose reuses external authorities by reference. Connectors reach the systems where work already happens — and everything resolves back to one graph, gated in CI.

Container diagram of the throughline ecosystem: the delivery team and an AI agent both author and propose items through throughline core and, on composed projects, through throughline-compose (the single command surface that gates the merged graph); compose builds on the core engine, resolves standard catalogues and auto-discovers connector resolvers; the Jira connector's core, bridge service and Forge app link Jira Cloud issues to nodes; the connect host mounts connectors; consumer repos and the TrueNAS platform (a worked adopter) are composed and gated in CI, grounding on the standard catalogues against git hosts.
The throughline ecosystem, container view. Core, composition, standard catalogues, the Jira connector (core, bridge service and issue-panel app), the shared connect host, consumer repos, the TrueNAS platform (a worked adopter grounding on ASVS), and the external systems they reach. Tap to open full size.
Why it matters

Nothing ungrounded ships — the build fails on an orphaned or circular graph.

Reuse standards, don't re-type them — pin an edition, and drift is surfaced rather than silent.

Governed AI — agents propose, humans ratify, and the reasoning stays on the record.

And it's just your repo — plain files under version control, no lock-in.

Contact
contact@timebacksolutions.co.uk

For questions about throughline. Time Back Solutions does not solicit or accept client engagements.