This page is machine-translated from French. Read the French original.

Glossary

Guest access

A limited, temporary, named access without an account, granted from the login page. The guest account and its conversations are deleted upon expiration. → Users, roles and organizations

Agent

An assistant authorized to act: call tools, chain steps, produce a concrete result. Triggered by @agent in a conversation, or automatically in Agent mode. → Agents Verification Status

Vector base

The database stores the vectors of document fragments and retrieves those closest to a question. → Vector base and embeddings

Skills library

A set of job instruction sheets, loaded on demand by an agent. This is distinct from skills, which involve executing code. → Skills library

PII Shield

The protection, disabled by default, replaces personal data with pseudonyms in the data sent to the language model, then restores the response. The agent tools receive the real values. → Security and confidentiality

Skill

A tool that an agent can call: web search, file reading, SQL query, calendar… Activated by an administrator for the entire instance. → SKILLS

Wiven Console

The Trust Functions screen: PII shield, sovereign mode, audit log, approved templates, walls, document permissions, installation health, license… Opened via a button next to the template selector. → Security and confidentiality

Context (window of)

The amount of text a template can process at one time. This includes the prompt system, pinned documents, retrieved passages, history, and query. When it is full, the oldest items are discarded.

Cutting

The splitting of a document into fragments of homogeneous size, with an overlap, to allow for searching.

Rights per document

A rule restricts access to a document to a list of people, including administrators. API keys are not subject to this rule. → Users, roles and organizations

Embedding

Transforming text into a vector—a list of numbers representing its meaning—is called vectorization. Two texts with similar meanings produce similar vectors. The interface refers to this as vectorization.

Workspace

The organizational unit of WivenLLM: a set of documents, settings, and conversations. Two spaces do not share their index or conversations, but they belong to the same instance: same storage, same administrators, same agent skills. → Workspaces

Pinned (document)

A document embedded in its entirety within the context of each message, without requiring a search. This is best suited for short and consistently relevant documents.

Thread

An independent conversation within a workspace. Threads share documents and workspace settings, but not their history.

Agent flow

A sequence of steps defined in advance and always executed in the same order. Unlike an agent, who decides their own path. → Agent flow

Fragment (chunk)

A piece of document resulting from the segmentation. This is the unit that the search finds and which is provided to the model.

Hallucination

A statement confidently produced by a model, but false or fabricated. Query mode and source citations are used to contain and detect it.

Audit log

The event log is copied to a hash string outside the database: any modification or deletion is visible. It is retained indefinitely by default. → Journals and supervision

LLM (Large Language Model)

The engine that writes the answers. WivenLLM is not one: it orchestrates the one you choose.

MCP (Model Context Protocol)

An open standard describing how an AI application discovers and uses external tools. → MCP Servers

Memory

A short, memorable fact about you, then used in conversations. → Memoirs

Agent Mode

The chat mode in which the assistant uses the tools himself, if the model and provider allow it, without going through @agent.

Chat Mode

The assistant uses your documents and its general knowledge. This is the default mode.

Query Mode

The assistant only responds based on your documents, and explicitly refuses when they do not contain the answer.

Multi-user mode

This mode enables named accounts, roles, invitations, organizations, and single sign-on. It is part of the WivenLLM Business offering.

Sovereign Mode

Outgoing connection control. By default, in Monitoring mode: it logs traffic without blocking. In Strict mode, it blocks anything leaving the company network that is not on the whitelist. → Security and confidentiality

Wall

A rule that closes off workspaces to certain individuals, including administrators, to separate conflicting mandates. → Users, roles and organizations

Organization

A grouping of users and spaces, with shared access and common login credentials. This is not isolation between companies. → Users, roles and organizations

Prompt system

The permanent instructions given to the model before each message: role, tone, format, prohibitions. The application's most effective tool. The interface refers to it as a system prompt. → Conversations

RAG (Research-Augmented Generation)

The central mechanism: find the relevant passages in your documents, provide them to the model, then generate the answer from them.

Model router

A mechanism that automatically directs each message to a template, based on rules regarding content, length, time, or the presence of an image. It has no confidentiality or cost requirements. → Model router

Second Brain

Your personal knowledge base, which spans all workspaces. A «second brain» within the interface. → Second Brain

Similarity threshold

The minimum degree of proximity a fragment must reach to be retained by the search. Too high, it excludes useful information; too low, it lets in noise.

Sources

The document excerpts displayed below an answer indicate what it is based on. They allow for verification.

Planned task

An agent instruction executed automatically according to a schedule. Available in single-user mode. → Planned tasks

Temperature

The degree of unpredictability of the responses. Low for factual answers, high for creative answers.

Web widget

A chat bubble to insert into a website, linked to a workspace. → Web widget