Skip to main content

​
Overview

In Qontext, a Context Vault is the core container where you bring together company knowledge scattered across many different tools. Qontext continuously structures and links that information into a living context graph that your AI workflows can query at runtime, so the AI gets relevant, up-to-date context without you manually setting it up for every new tool or use case. Practically, a Context Vault is where you:
  • Define the scope: Company-wide, team-specific, or use-case-specific.
  • Connect and sync data sources: Updates in the source are reflected in the vault.
  • Retrieve context: Via API, MCP, or workflows (n8n, Make, Zapier, custom agents), using one vault as the shared context foundation.
Each vault belongs to exactly one workspace; each workspace can have multiple vaults.

​
Usage

​
How to set up a Context Vault in the platform

1
Sign in to the Qontext app and create a new Context Vault.Add vaultGive it a name that reflects its scope.Create new vault
2
(Optional) Configure ingestion and retrieval instructions at the vault level. For more information, see System Instructions.
3
Create an API key to ingest and retrieve any data
4
Ingest the first data into the vault (see Data Sources for details).
5
Link your AI tools (see Context Access for details) to Qontext to ensure the generated output is always up-to-date and fully aware of the important company knowledge.

​
Behind the scenes: how vaults are built and updated

When you ingest data into a vault, Qontext automatically organizes it into context-rich relationships and turns it into a knowledge graph so retrieval can be relationship-aware and relevance-ranked. This intelligence layer makes Qontext retrieval smarter by design, and the process runs entirely in the background with no setup required. How it works

Entity extraction

Detects key entities like people, companies, and concepts.

Relationship mapping

Connects related entities across your data sources.

Graph construction

Builds a knowledge graph that reflects real-world relationships.

Continuous updates

Keeps your context current as new data is ingested.
Why it matters This structure is what enables Qontext to go beyond traditional RAG systems.
  • Reflects relationships, not just keywords.
  • Preserves semantic meaning across tools.
  • Stays accurate and current over time.
  • Enables more intelligent context retrieval.
You don’t need to configure any of this; the pipeline runs automatically for every vault.

​
Best practices

Prefer one vault when possible.A unified vault provides the AI a holistic view, but separate vaults are necessary for strict access control, varied instructions, or compliance-driven data isolation. Prioritize a centralized structure unless security requirements or team-specific scopes demand fragmented environments.
Personal vs Organization workspace.Choose your personal workspace for individual use cases and private context that remains isolated from the team.Use your organization’s workspace to centralize shared data sources, manage collaborative API keys, and maintain governed context across team workflows.

​
FAQ

While you can create multiple vaults for different teams or use cases, we recommend a holistic approach with fewer, larger vaults to give your AI a unified perspective. Create separate vaults only when sensitive data requires strict isolation or when a specific use case demands completely different system instructions.
Data is ingested into a specific vault. To “move” context, you would ingest the same or updated content into another vault; there is no built-in copy-between-vaults operation.