Skip to content ReleaseVezual Beta is in the Plugins Console

MCP

Your AI client.
Your database, in reach.

MCP (Model Context Protocol) lets an AI assistant use tools in other software. See how a database connection works, then check server availability for your setup.

MODEL CONTEXT PROTOCOL
What tables are available?
AI clientYour workspace
OriginChain MCPFive database tools
Your databaseOriginChainDB
oc_list_schemasExample result
customersid · name · email
ordersid · customer_id · total
Your client calls a tool. OriginChainDB returns the data.

Compatibility preview · request availability

A first request

From a question
to a database result.

Start with the structure of your data. Then choose the query you need.

01 / CONNECTION

Choose the database
your client can use.

Ask our team to confirm a server version that supports your client and database connection. Use the configuration supplied for that version.

This walkthrough illustrates a local server sending database requests over HTTPS. The server for the new access-token connection is not yet available as a verified public install.

Connection checklistPreview
Client
A supported MCP client
Server
A compatible, verified version
Access
Credentials for that server's connection format

Use the setup instructions supplied for your client.

02 / TOOL CALL

See what the client
is asking for.

For this question, the client can use oc_list_schemas. Review the tool name and arguments in your client before continuing.

The review controls depend on your MCP client and its settings.

QuestionExample

“What tables are available?”

Tooloc_list_schemas

Inspect the tables, columns and indexes in the configured tenant.

DestinationYour OriginChainDB instance
03 / RESULT

Read the schema.
Ask the next question.

The returned schema describes your tables and columns. Your client can use these names to prepare a query.

Review returned data and errors before using the answer in your work.

Schema resultIllustrative
Example tables and columns returned by schema discovery
TableColumns
customersid, name, email
ordersid, customer_id, total
A next question

“Show the orders for a customer.”

Illustrative walkthrough. No database connection or query is made.

Tool directory

Database tools.
A clear job for each.

These are tools from the earlier integration. The available list must be confirmed for the server version and instance you use.

01 oc_list_schemas

Understand the structure

Inspect your tables, columns and indexes.

Example intent“What tables are available?”

02 oc_sql

Run SQL

Run supported SQL with parameters.

Example intent“Which orders were placed today?”

03 oc_ask

Ask in plain language

Ask a question about your tables.

Example intent“Show new customers this month.”

04 oc_vector_topk

Find similar records

Find similar records using a vector (a numeric representation of data) and optional filters.

Example intent“Find records similar to this example.”

05 oc_fts_search

Search indexed text

Find text using words and phrases.

Example intent“Find passages that mention delivery.”

Ask and search tools require the corresponding instance capabilities. Multiple tool calls remain separate requests.

Setup & access

A local connection.
Your database access.

MCP gives your client a set of database tools. The connection still uses your configured credentials.

Explore authentication
01

Runs alongside your client

Your client starts a local server and talks to it through stdio (standard input/output). The server sends database requests over HTTPS.

02

Uses the access you configure

Your credentials and the database determine which requests can run. Verify the intended access before connecting an AI client.

03

Get the right setup

A verified public server for the new connection flow is still pending. Our team can confirm availability and provide the correct setup when your configuration is supported.

Request MCP access

Build with OriginChainDB

Connect your client to your data.