The Veyo Portal is where you set up projects, connect repositories, monitor ingestion, manage credentials, and control team access. All customer-facing interaction happens here.
The Portal is the human interface to Veyo. Set up once, then let your AI assistants query the data via MCP.
Step-by-step project creation: name your project, connect source and target repositories, assign credentials, and trigger your first ingestion.
Each project supports multiple repositories with distinct roles: source version, target version, and custom extensions. Compare any combination.
Connect GitHub and Azure DevOps repositories using encrypted PATs. Credential resolution cascades: repository, project, then user scope.
Watch ingestion progress live via Server-Sent Events. Track job status, file counts, error rates, and batch completion across all your projects.
Generate, rotate, and revoke API keys scoped to specific projects. Track usage with last-used timestamps and request counts.
Invite team members and assign roles per project: Owner, Admin, Member, or Viewer. Control who can ingest, query, and manage settings.
Name your project and choose your organisation. The wizard guides you through each step with validation and helpful defaults.
Point to your GitHub or Azure DevOps repos. Assign each as source (current version) or target (upgrade version). Add credentials — encrypted at rest with Fernet.
Trigger ingestion and watch progress in real-time. The system detects changes per-commit, downloads only modified files, and parses them through 20+ extractors.
See every ingestion job in real-time: pending, downloading, processing, completed, or failed. Jobs stream updates via SSE so the dashboard is always current. Filter by project, status, or batch.
Each project card shows connected repositories, last ingestion timestamp, file counts, and active API keys. Drill into any project for full configuration and history.
PATs and API tokens are encrypted with Fernet symmetric encryption. Credentials are scoped and resolved hierarchically: repository > project > user. Rotate or revoke at any time.
Every user action is logged: project creation, ingestion triggers, credential changes, and access modifications. Audit logs are immutable and queryable for compliance.
Organisation-level system repositories make standard Business Central base application code available to all projects. Set up once, share across your team.
Veyo tracks the head commit of each connected repository. When code changes, it detects modified files via manifests and creates targeted ingestion jobs — only processing what changed.
Every project gets a unique collection prefix that namespaces all 12 semantic tables. Project A cannot see Project B's data — even within the same organisation.
Organisations group users and projects under a single billing and management boundary. Members can be assigned different roles per project within the org.
System repositories (shared BC base application code) are scoped to the organisation level, so all projects in an org can reference the same standard code without duplicating ingestion.
In cloud mode, Clerk handles the full authentication lifecycle: sign-up, sign-in, session management, organisation membership, and webhook-driven data sync.
In on-premise mode, the Portal derives the JWKS endpoint from the publishable key and verifies tokens locally. No Clerk API calls needed after initial key set fetch — fully offline-capable.
Either way, the Portal experience is identical. Users sign in through Clerk's hosted components, and the Orchestrator validates every request.
Connect your repositories, generate an API key, and let your AI tools query the real data.