Self-hosted · Air-gapped by design

The enterprise AI platform that never leaves your network.

Mentior runs assistants, autonomous agents, and document knowledge bases entirely inside your own infrastructure. Local models, zero external calls — your data simply never goes anywhere.

Early access for teams running AI on-prem. No spam, just the launch.

No data egress
Runs fully on-prem
Local LLM inference
Role-based access
The platform

Everything an AI team needs, behind your own firewall.

Build, connect, and observe AI systems without a single byte leaving your perimeter.

Complete network isolation

The whole system operates inside the customer's own network with no external communication. No telemetry, no API round-trips, no vendor in the data path — verifiable air-gapped operation that satisfies the strictest compliance and data-residency requirements.

RAG knowledge bases

Ingest PDFs, text, and markdown into local vector search. Assistants answer from your documents, with context grounded in your sources.

Configurable assistants

Define a system prompt, pick a default model, and link knowledge bases. Sessions can override the model per chat for flexibility.

Visual agent workflows

Drag-and-drop canvas to chain assistants, agents, knowledge, and tools. Run them with live, step-by-step node status streamed in real time.

Streaming chat

Token-by-token responses over WebSocket, with tool-execution logs and RAG context surfaced inline as the answer is generated.

Built-in LLM observability

Every prompt, completion, tool call, and latency is traced and stored locally. See exactly what your models did and why — full auditability without shipping logs to a third party.

Extensible tool system

Connect agents to filesystem, web search, and database tools through a pluggable registry — safe, scoped, and easy to extend.

How it works

From your documents to grounded answers — without leaving home.

01

Deploy in your network

One stack runs everything — database, vector store, local model inference, and observability — inside your own infrastructure.

02

Load your knowledge

Upload documents into knowledge bases. They're chunked, embedded, and indexed for fast local retrieval.

03

Configure assistants & agents

Wire prompts, models, knowledge, and tools together — visually for workflows, or directly for chat.

04

Run, observe, iterate

Teams chat and run workflows while every step is traced locally for full visibility and control.

Bring AI to your data —
not your data to AI.

Join the waitlist for early access to the self-hosted Mentior platform.