Introduction

Reloqui is a multi-tenant release-intelligence platform. It answers a deceptively hard question that every engineering organization struggles with:

What actually shipped in this release — where, when, and can I trust that answer?

The problem

The evidence for any release is scattered across half a dozen systems: pull requests in GitHub or GitLab, issues in Jira or Linear, CI builds, published artifacts and image digests, deployments from your delivery tooling, and infrastructure changes from Terraform. Answering “what’s in v2.3.0 of the checkout service, and is it safe to publish?” usually means a human stitching those systems together by hand — slowly, and without a way to prove the answer is complete.

What Reloqui does

Reloqui ingests those signals as typed evidence, then runs a deterministic-first, heuristic-second reconciliation engine that produces one explainable answer per release:

• What’s included — every PR, issue, build, artifact and deployment attributed to the release, each with a reason.
• A confidence score (0–1) and a tier (high / medium / low) that grades how well-supported the answer is.
• Exceptions — orphaned changes, uncertain items, force-pushes and backports surfaced instead of hidden.
• Audience-specific release notes generated from that evidence, where every claim is backed by a stored record.

Around that core sit governance (policies, approvals, freeze windows), a service catalog with a dependency graph, twelve source connectors, 360° views, notifications, and a full admin/enterprise surface (SSO, SCIM, tokens, audit).

Why “evidence-backed” matters

Reloqui never asks you to trust a number on faith. Every release exposes the exact evidence behind its score, the classification reason for each item, and a frozen snapshot at publish time. That makes release truth auditable — you can hand it to a compliance reviewer, an executive, or an on-call engineer and each sees a view they can rely on.

How it works, in one picture

 SCM · Trackers · CI/CD · Registries · Deploys · Terraform · Incidents
                          │  (webhooks + polling)
                          ▼
            Typed canonical evidence  (scm.pr.merged, build.completed, …)
                          │
                          ▼
     11-stage reconciliation engine  ── scope → collect → normalize →
     deterministic attribution → heuristic attribution → classify →
     backport/force-push → confidence scoring → policy → snapshot
                          │
                          ▼
   Release truth:  included evidence · confidence + tier · exceptions ·
                   audience notes · 360° views · governance decisions

Who it’s for

• Release managers who need a trustworthy publish decision.
• Engineers who want an explainable list of what’s in a release.
• Executives who want portfolio-level delivery visibility.
• Compliance & auditors who need provable, frozen release records.
• Platform teams who want one catalog and dependency graph across services.

Next steps

• Follow the Quick start — three scenarios that get you productive fast.
• Or open the live demo and click through a fully-seeded workspace, no login required.