Postgres database for AI agents
Every workspace gets a real Postgres database. Your agent creates tables, runs SQL, loads CSVs and reads its own data back, with no provisioning step and no connection string to hand out.
Copy the setup prompt.
Your agent does the rest.
Fetch and execute the appropriate instructions to set me up for superagnt_ from https://mcp.superagnt.com/agent-setup/prompt.md» On by default. The agnt_db_* tools are part of the base tool surface, so a fresh MCP connection already has them.
Everything your agent can do with it
agnt_db_list_tables returns tables, columns, types and row-level security status, so the agent reads the schema instead of guessing at it.
agnt_db_apply_migration runs DDL and records it, so the schema an agent built is auditable and repeatable.
select, insert, upsert, update and delete with filters, so simple work needs no hand-written SQL. Update and delete both require a filter, which is what stops a table-wide accident.
agnt_db_execute_sql for anything the helpers do not cover, with a read-only mode when you want a statement that cannot write.
Stream a CSV from workspace files into a table server-side, with type inference, batched upsert and dedup on conflict.
agnt_db_status reports provisioning state and readiness, so an agent waits rather than fails when a database is still coming up.
List installed Postgres extensions, and every migration applied through the tool.
tool names12as your agent sees them
agnt_db_statusagnt_db_list_tablesagnt_db_list_extensionsagnt_db_list_migrationsagnt_db_apply_migrationagnt_db_execute_sqlagnt_db_selectagnt_db_insertagnt_db_upsertagnt_db_updateagnt_db_deleteagnt_db_load_csvThree jobs it can start on today
Give an agent somewhere to keep its work
Instead of re-deriving state on every run, the agent creates a table on first use and writes what it found, so the next run starts from what the last one learned.
Pipeline output the team can query
A scraping or enrichment job writes rows as it goes. Analysts query the same database directly, so nothing important lives only in a chat transcript.
CSV in, structured out
Someone uploads a list. The agent loads it with agnt_db_load_csv, enriches each row, and writes the results back to the same table with an upsert.
What comes back
Your agent reads this response directly, with no parsing layer of your own in between.
https://mcp.superagnt.com/mcptool callagnt_db_select{
"table": "leads",
"columns": ["id", "company", "status"],
"filters": { "status": "qualified" },
"order": "created_at.desc",
"limit": 3
}{
"rows": [
{ "id": 4182, "company": "Meridian Analytics", "status": "qualified" },
{ "id": 4179, "company": "Halcyon Freight", "status": "qualified" },
{ "id": 4171, "company": "Ridgeway Dental", "status": "qualified" }
],
"row_count": 3
}What it costs to run
Platform tools carry no per-tool price. What you pay is the AI and credit usage the work causes, on the same workspace balance every other tool draws from.
Each run is itemized, so you can see which agent spent what and on which call.
how credits work →Before you wire it up
01Do I have to provision it?
No. The database exists per workspace. agnt_db_status reports whether it is provisioning, active or failed, so an agent can wait for readiness rather than error out.
02Can the agent break the schema?
It can change it deliberately, through agnt_db_apply_migration, and every migration is recorded and listable. Update and delete both require a filter, so a table-wide write is not something a single call does by accident.
03Is it real Postgres?
Yes. Extensions, indexes, constraints and arbitrary SQL all work. agnt_db_execute_sql takes a read-only flag when you want to guarantee a statement cannot write.
04How do I get these tools?
Connect your client to https://mcp.superagnt.com/mcp. The database tools are in the default surface, so no agnt_tools_enable call is needed.
Point your client at https://mcp.superagnt.com/mcp and your agent has this, plus everything else on one balance.