Find a table before querying it
Question: Which table holds CMBS loan-period observations, and what does one row represent?
Reveal MCP serves catalog metadata for a provisioned account. Request access and ask for the catalog scope and MCP client configuration. There is no public connection URL or self-serve signup in this guide. The lookup below reads metadata; it does not query loan balances.
1. Look up the known table
In your configured MCP client, call cmdrvl.com.metadata.lookup with these JSON arguments:
{"reference": "PROPERTY_MART.LOAN_PERIOD_FACT","limit": 3}
The reference is a known table name, not a free-text question. The limit bounds catalog matches; it is not a count of loans or table rows.
2. Read the catalog response
In an authenticated read-only check on November 3, 2026, the lookup returned found: true with one result. Its catalog key was snowflake://edgar_db.property_mart/loan_period_fact, and the physical table was EDGAR_DB.PROPERTY_MART.LOAN_PERIOD_FACT. The recorded grain was one observation per CIK, asset number and reporting-period end date. The response also said has_query_contract: false.
The grain matters: counting rows would count reporting periods, not distinct loans. has_query_contract: false does not mean the table is unusable. It means this lookup did not supply a query contract. Confirm the live columns and your account's row access before composing SQL.
This step contributes source meaning and a recorded source pointer. It does not cite a filing, verify a balance or prove that every relevant table is cataloged.
If the lookup does not match
- Check the reference spelling and the catalog scope provisioned for your account. A missing catalog entry is not proof that the source data does not exist.
- Read any applied-filter or dropped-filter caveat before treating a broader search as an exact match. A broad topic search may leave a filter unapplied.
- If several tables match, compare their physical names, descriptions and grains before choosing one. Do not select by catalog count alone.
Next, describe and query this table with Data MCP. For the product's coverage and access limits, see Reveal MCP.