The full catalogue
Every module in Vigil
24 modules across 7 categories, under a single multi-tenant, role-based login. Pick a module. Every lifecycle below is the real one, straight out of the product.
Built in, no add-ons
AI layer & MCP
Intelligence & ReportingOne pluggable, OpenAI-compatible provider behind every AI feature in Vigil, and Vigil itself exposed as an MCP server so any LLM can work with your data.
- Drafts helpdesk replies, summarizes threads, parses and scores CVs, answers from your own documents, writes meeting minutes and analyses vigilance frames
- Swapping vendors is a config edit, not a code change; self-hosted models work through the same interface
- Embeddings are optional: semantic search degrades to full-text instead of breaking
- The MCP server exposes tickets, attendance, knowledge, tasks and more as tools to Claude, ChatGPT, Copilot or any MCP client
- Every AI call is scoped to the organization and role of the person who triggered it
Real lifecycleMany sources, one engine
One provider in the middle, every module around it
Modules never talk to a vendor directly. They call the AI layer, which holds the provider config, the prompts and the optional embeddings model. The same layer runs the other way as an MCP server: an external LLM connects, sees Vigil as a set of tools, and works with your tickets, attendance and knowledge under the caller’s permissions.
One interface ›››
››› Returns
HelpdeskThread to draft
Knowledge baseQuestion + your docs
RecruitmentCV text
Scrum meetingsTranscript
Shift VigilanceFrames on demand
MCP clientClaude, ChatGPT, Copilot
Drafted replyAgent reviews, then sends
Answer with sourcesGrounded in your docs
Parsed profile & scoreOptional JD-fit
Meeting minutesDecisions and actions
Tool resultsScoped to the caller’s role
AI layerProvider · Prompts · Embeddings