- LIVE DATA
- SYNCED
- DISCOVERED
- HUMAN
Short answer
How can I keep my AI's knowledge up to date?
You trigger your agent to reach out to remote SQL databases or cloud storage to pull in new information. This keeps its context grounded in your live data. You'll see how the ingestion and discovery process works below.
The results of ingestion.
Where your data lands.
Once the connection is live, your agent handles the heavy lifting of data retrieval.
LIVE DATA
Fresh context
The agent pulls the latest records from your remote sources. You get answers based on current facts, not stale files.
SYNCED
Automated updates
Instead of manual uploads, the agent fetches new entries from your storage. This keeps the knowledge base consistent.
DISCOVERED
Source visibility
Your agent identifies available data locations. It knows exactly where to look for the information you need.
HUMAN
Manual oversight
You decide which sources to pull from and when to trigger a refresh. The AI won't change your source permissions.
The workflow
What your AI does when you request a data refresh.
The agent moves from identifying where data lives to pulling it into its active context.
Locate sources
The agent scans your connected environments to find available databases or storage buckets.
list_data_sourcesPull new data
The agent reaches into the identified source to grab the specific records you requested.
ingest_dataVerify connection
The agent confirms it can successfully read the target files or tables before attempting a full pull.
list_data_sourcesUpdate context
The agent integrates the newly fetched information into its immediate working memory.
ingest_data
Try it
Copy these to start.
Use these prompts to trigger the data ingestion process.
Starting points
Replace the bracketed text with your actual database names or cloud bucket paths.
Cognita Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_SHxG8KJRFQw2m3lX2GfeLS7zRF34WrsRFB0j95Nv/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — Cognita capabilities are ready to use.
{
"mcpServers": {
"cognita-rag-framework-mcp": {
"url": "https://edge.vinkius.com/vk_preview_SHxG8KJRFQw2m3lX2GfeLS7zRF34WrsRFB0j95Nv/mcp"
}
}
}List all my available data sources so I can see what you can access.
Ingest the latest customer records from my production SQL database.
Check my cloud storage bucket for any new CSV files added today.
Update your knowledge by pulling data from the sales reporting table.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect Cognita once, then ask.
You only need to link your credentials once. Vinkius keeps your logins encrypted and ready for your agent to use.
Connect Cognita to your AIFAQ
How the data ingestion works.
- 01
Can the AI delete data from my SQL database?
No. The agent only has permission to read and ingest data, not to modify or delete your original records.
- 02
Does the AI need me to upload files manually?
No. Once connected, the agent uses the ingestion tools to pull data directly from your remote sources.
- 03
How does the agent know which data sources are available?
The agent uses the discovery tool to list all connected sources it has permission to access.
- 04
Can the AI change my cloud storage permissions?
No. The agent cannot change security settings or access levels on your remote storage.
- 05
Will the AI automatically pull data every hour?
No. The agent pulls data when you specifically ask it to or when you trigger an ingestion command.
More questions about Cognita? The Connector page answers them. See everything the Cognita Connector can do
Connect Cognita to Claude, Cursor, ChatGPT & more
