Introduction
Premora is an on-prem enterprise knowledge platform. It breaks down knowledge silos and gives engineers — and the coding agents working alongside them — reliable, permissioned context about the work they are doing.
These docs are written for the teams who install, administer, and use Premora inside their own environment.
What Premora does
Premora ingests from the systems a team already uses, preserves the source ACLs through the entire pipeline, and surfaces that knowledge four ways:
- ACL-aware search — retrieval that only ever returns what the asking user is permitted to see, with citations and source lineage on every result.
- A deterministically generated wiki — raw source content materialized into structured, readable pages with full source lineage.
- A governed research flow — conversational, evidence-backed work surfaced through the
AskandReviewexperiences. It combines search, governed read-only data questions, and optional web research into cited answers, then lets users promote durable findings into the shared wiki under review. - An agent skill — a Claude Code / Codex skill that feeds grounded, cited internal context into AI-assisted coding.
Success looks like: an engineer asks a question about an incident, a system, or a decision, and gets a grounded answer in under 30 seconds with traceable citations — not a hallucination.
Deployment shape
One Premora deployment is scoped to a single enterprise customer. There is no SaaS-style cross-customer multitenancy; partitioning for teams and workspaces happens inside that one deployment. The platform runs fully on premises or in a customer-controlled private cloud, and the core read path keeps working without any external SaaS dependency.
Premora Core installs as a single-box Docker Compose stack via the premora CLI, from signed
and versioned artifacts — no Kubernetes to operate. See Requirements
and Installation to get started.
The product family
| Product | What it is |
|---|---|
| Premora Core | The single-process (omnibus) application platform: gateway, identity, connectors, ingest, knowledge, search, governed research flow, audit. |
| Premora Inference | The optional host-managed GPU runtime that fronts vLLM for private on-prem inference. |
| Premora Installer | The customer-facing premora CLI that installs Core and Inference from connected or offline (USB) artifacts. |
| Premora Skill | The Claude Code / Codex skill package that lets coding agents query the knowledge base. |
Where to go next
- Requirements — what you need before installing.
- Installation — connected, registry, or air-gapped.
- Integration catalog — what Premora connects to, and how to add a source.
- Connect a coding agent — use Premora as a private model endpoint.
- Administration — identity, access control, config.