Multi-Database AI Query Agent
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MAX_ROWS | No | Default and maximum requested query rows; hard maximum 1000. | 100 |
| DATABASES_JSON | No | JSON array of {id, description, url, schemas} database configurations. | [] |
| MAX_CELL_CHARS | No | String/cell truncation limit; maximum 10000. | 4000 |
| ANTHROPIC_MODEL | No | Configurable Claude model ID. | claude-sonnet-4-6 |
| MAX_AGENT_TURNS | No | Claude round-trip budget; maximum 30. | 12 |
| ANTHROPIC_API_KEY | No | Secret API key for Claude; optional for diagnostics/tests, required for natural language. | |
| HR_ADMIN_PASSWORD | No | Compose initialization password for the HR admin user. | |
| HR_READER_PASSWORD | No | Password for the separate HR read-only reader. | |
| DATABASE_CONFIG_FILE | No | Optional local JSON array file; takes precedence over DATABASES_JSON. | |
| STATEMENT_TIMEOUT_MS | No | Per-statement timeout in milliseconds; allowed 10–30000. | 3000 |
| ECOMMERCE_ADMIN_PASSWORD | No | Compose initialization password for the ecommerce admin user. | |
| ECOMMERCE_READER_PASSWORD | No | Password for the separate ecommerce read-only reader. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_databasesA | List configured database identifiers, descriptions and availability; no credentials. |
| list_tablesC | Discover authorized base tables, comments and approximate row counts. |
| get_schemaB | Inspect columns, keys, foreign-key relationships and indexes of discovered tables. |
| run_queryB | Execute one validated read-only SELECT/CTE, returning bounded structured results. |
| explain_queryB | Describe SQL structure and PostgreSQL EXPLAIN estimates, without ANALYZE. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 5 tools
Each tool maps to a distinct stage of a read-only analytics workflow: database discovery, table discovery, schema introspection, execution, and plan explanation. The only superficially similar pair (run_query vs explain_query) is clearly separated by 'returns bounded structured results' versus 'EXPLAIN estimates, without ANALYZE'. An agent can pick the right tool without guessing.
All names are snake_case verb_noun pairs (list_*, get_*, run_*, explain_*) with a consistent, predictable structure. The verbs accurately signal the operation type, and there are no camelCase or style outliers.
Five tools is tightly scoped for a read-only query agent and every tool earns its place in the discovery→introspection→execution chain. Nothing appears redundant or bolted on.
The core lifecycle for a read-only agent (discover → inspect → execute → explain) is fully covered, and the absence of write tools is clearly intentional. Minor gaps remain: no cross-database/column search or namespace listing, so an agent must walk databases/tables sequentially to locate a field.