Product evidence
Clavinci is a commercial, company-operated engineering intelligence product. It joins structured AI-tool activity to code, delivery, review, and stability outcomes without collecting the work itself.
Evidence checked: 2026-07-16
01 / evidence model
Clavinci starts with assistant sessions, token classes, models, cache behavior, and cost. It attributes those sessions to commits and Jira work, then follows the relationship through pull requests, CI, deployments, incidents, and 30-day code survival.
01.1 / operating model
Collectors run near the engineering tools and retain local history. The company operates the hosted API, database, dashboard, upgrades, backups, and service health. Customers control connector authorization, project scope, identity, retention requirements, and permitted workforce use.
02 / inputs
Each source contributes structured evidence. Content that is not required for measurement stays outside the event contracts.
Session boundaries, tools, models, token classes, cache reads and creates, timestamps, project context, tool events, and compaction metadata.
Commit identity, repository and branch context, file paths, diff statistics, line deltas, parent relationships, and deployment markers.
Jira issue keys, types, status transitions, priority, assignee, story points, sprint membership, epics, and lifecycle timestamps.
Pull request events, review state, CI timing and outcomes, deployments, hotfixes, incidents, on-call pages, and restoration timing.
03 / joins
Time proximity, file overlap, explicit session trailers, and Jira matches produce proportional links. Attribution versions can be replayed, disputed by engineers, and resolved by an admin.
Issue keys in commits, branches, and pull requests connect code changes to Jira issues, sprints, epics, story points, spillover, aging work, and time in status.
Commit SHA and pull request identity connect attributed work to review timing, CI, deployments, DORA, rollback, rework, and 30-day survival.
Incident adapters connect changes to hotfixes, change failure, incident rate, mean time between incidents, on-call load, reopen rate, and restoration time.
04 / outcomes
The product measures cost, delivery, review, stability, quality, and adoption from the same attributed cohorts.
Actual and API-equivalent cost per issue, story point, CI run, and merged line that survives 30 days
↗DORA, AI lead-time delta, change failure, rollback, rework, and durable code survival
↗Sprint velocity, spillover, cycle count, time in status, aging work, stale issues, and epic economics
↗Pull request size, first review, merge after approval, iterations, self-merge, reviewer turnaround, and review load
↗Hotfixes, incidents, mean time between incidents, on-call page rate, reopen rate, and restoration time
↗Re-rolls, suggestion acceptance, tool-task fit, prompt complexity, churn, compaction, adoption, and tool mix
↗05 / boundary
06 / integrations
07 / implementation
The checked-in product contains nine AI collector implementations, git collection, replayable attribution, pricing and cost calculation, Jira and incident adapters, DORA, sprint, PR, stability and AI-quality metric packages, local and cloud APIs, governed dashboard surfaces, notifications, and collection-health aggregation.
08 / limitations
Clavinci does not offer customer-managed, self-hosted, or on-premise deployment.
Jira is the current work-tracker integration. Linear, Asana, Shortcut, and GitHub Issues are not current product integrations.
Clavinci does not provide developer surveys, industry benchmark programs, generalized BI, portfolio finance workflows, AI code review, or pull request automation.
The product measures structured metadata. Its content boundary intentionally rules out content-aware coding-agent context and code analysis.
Public prices, plan boundaries, and service commitments are not final and must not be inferred from the implementation status.
Evidence checked: 2026-07-16
Clavinci is built to make the full engineering outcome path inspectable while keeping the work itself outside the measurement system.