Trust

How Threader is built. The architecture, the data handling, the testing, and the people behind the four products. Every line of code is reviewed, tested, and understood. This page is the detail behind that claim.

A small UK-based team

Threader is built and maintained by a small UK team. The technology is AI-assisted, human-led: AI is used as a typing assistant, not as a substitute for engineering judgement. Every architectural decision is the team’s. Every test is reviewed.

We do not outsource our software development.

Isolated workers, structured data, deliberate architecture

Each Threader product runs on isolated worker infrastructure. A fault in one tool cannot cascade into another. If one component fails, the others keep running. Every cascade output is stored as structured JSON, versioned on save, so no work is silently lost or overwritten.

Data model changes follow a structured migration process. Breaking changes are versioned. Nothing ships to users without a migration path for existing data.

Named infrastructure, encrypted in transit and at rest

Threader runs on Cloudflare Workers for application logic, Supabase for structured data, and the Anthropic API for AI inference. All data is encrypted in transit (TLS) and at rest. We name our infrastructure providers because specificity is the trust signal; vagueness is not.

Customer data is never used to train AI models. Anthropic’s API does not retain prompts or completions for training under our usage agreement.

Real outputs, verified against expected structures

Every release runs through automated integration tests that verify real outputs against expected structures. We test the paths users actually follow, not synthetic coverage numbers. Tests run before any code reaches production. Bugs found post-release are logged in the per-product changelogs alongside the fixes.

Your data is yours, and you can take it

Full data export is available at any time from within each product. Outputs are exportable as structured files. Account deletion is straightforward and removes all associated data within 30 days, in line with UK data protection law.

Public artefacts, not marketing claims

The platform status page is public at threadertechnologies.com/status. The macro context layer that grounds every output is exposed at macro-context.json. Per-product changelogs are open on each product’s support hub. You do not have to take our word for any of this.