polydb

PolyDB: one API key, every data shape your agent needs PolyDB is a hosted multi-model database made for AI agents.

CommunityOnlineAPI key requiredGlobalFreeCan modify data

What it can do

  • Query Sql: Execute raw SQL query on PolyDB. Only read-only queries are permitted (SELECT, WITH, EXPLAIN, SHOW). Write operations (INSERT, UPDATE, DELETE, DROP, etc.) are rejected. Args: query: SQL que
  • List Tables: List tables in your agent database, or describe one table's columns in detail. This introspects the AGENT database live (not a hardcoded list) and never shows PolyDB's own internal `polyd
  • Store Document: Store JSON document in a collection. Args: collection: Collection name document: JSON document to store document_id: Optional document ID (generated if not provided) database: Optional

What data it sees

Do you need an account

An API key from the service settings is required

PolyDB: one API key, every data shape your agent needs

PolyDB is a hosted multi-model database made for AI agents. One API key gives your agent a private, isolated database that holds many kinds of data at once. You don't need to set up or connect a separate service for each kind.

What your agent can store and query

  • Memory and knowledge: session-scoped memories with semantic recall (store_memory / recall_memory), plus tenant-wide knowledge learned once and reused (store_knowledge / search_knowledge)
  • Vectors for RAG: store_vector / search_vectors, with collections created on first insert
  • Documents and SQL: JSON documents, plus read-only SQL over your tables (query_sql, which accepts SELECT, WITH, EXPLAIN and SHOW)
  • Key-value with TTL
  • Graphs: nodes, edges and graph queries
  • Time-series and temporal: metrics over time, and time-travel queries on versioned entities
  • Geospatial: nearby and bounding-box search
  • Full-text search
  • Files: blobs and S3-compatible object storage
  • Streams: publish and consume

Made for agents

  • 69 tools with typed input and output schemas. search_schema returns the catalog grouped by data type.
  • Multiple named databases per account. Pass database on any call to pick one.
  • Stable error codes. Auth errors include how_to_fix and ready-to-paste setup snippets, so an agent can tell a human exactly what to fix.
  • Tenant isolation. Every request is scoped to the account that owns the key, and you can't see other tenants' data.

Getting started: get an API key at https://beta.polydb.dev/signup and paste it into the polydbApiKey field. Docs: https://beta.polydb.dev/docs

How to connect

How to connect polydb to Claude

The server address to paste into Claude settings:

https://polydb--richard-white.run.tools

  1. Open Claude (claude.ai or the desktop app).
  2. Go to Settings → Connectors.
  3. Click Add custom connector.
  4. Paste the server address copied below into Remote MCP server URL and click Add.
  5. In a chat, click + → Connectors and switch the new connector on.

Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

Authorization

After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

How to connect polydb to ChatGPT

The server address to paste into ChatGPT settings:

https://polydb--richard-white.run.tools

  1. Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
  2. Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
  3. Open Settings → Connectors and click Create.
  4. Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
  5. Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
  6. In a new chat open + → Apps/Connectors and enable the connector.

OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

Authorization

On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

How to connect polydb to Cursor

The server address to paste into Cursor settings:

https://polydb--richard-white.run.tools

Fastest: click Open in Cursor below and confirm the prompt.

Manually:

  1. In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
  2. Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
  3. Save the file. The server appears in the MCP list; authorize it there if asked.

Authorization

If the service needs sign-in, an authorize button appears next to the server in the MCP list.

Server tool list (60)

Raw names from tools/list. Only developers need these.

query_sql Execute raw SQL query on PolyDB. Only read-only queries are permitted (SELECT, WITH, EXPLAIN, SHOW). Write operations (INSERT, UPDATE, DELETE, DROP, etc.) are rejected. Args: query: SQL query to execute (read-only) params: Optional query parameters database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with query results (rows, columns, count)
list_tables List tables in your agent database, or describe one table's columns in detail. This introspects the AGENT database live (not a hardcoded list) and never shows PolyDB's own internal `polydb_*` infrastructure tables. Use it to discover what tables you (the agent) have created via `create_table`, then read them with `query_sql`. Pass `table` to get full column detail for that one table instead of the table listing (this absorbs what would otherwise be a separate `describe_table` tool). Args: table: Optional table name. If given, returns column detail for that table instead of the table listing. database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Without `table`: Dict with {"tables": [str, ...], "count": int}. With `table`: Dict with {"table": str, "columns": [{name, type, nullable, default, primary_key, position}, ...]}.
store_document Store JSON document in a collection. Args: collection: Collection name document: JSON document to store document_id: Optional document ID (generated if not provided) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with document_id
search_documents Query documents in a collection. Args: collection: Collection name filter: JSON filter query limit: Maximum number of results database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with matching documents
get_document Get document by ID. Args: collection: Collection name document_id: Document ID database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with document or error if not found
delete_document Delete document. Args: collection: Collection name document_id: Document ID database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with success status
store_vector Store embedding vector with metadata. Args: collection: Vector collection name vector: Embedding vector (array of floats) vector_id: Optional vector ID (generated if not provided) metadata: Optional metadata database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with vector_id
search_vectors Semantic similarity search on vector embeddings. Args: collection: Vector collection name query_vector: Query vector for similarity search k: Number of nearest neighbors to return metadata_filter: Optional dict of metadata equality constraints, e.g. {"category": "docs"}. Only vectors whose metadata matches all pairs are returned. On the cloud (pgvector) and SQLite strategies the filter is applied before the top-k cut (pre-limit, exact). On the embedded FAISS strategy it is applied over the top 10×k nearest candidates (an approximation — a match ranked beyond the 10×k nearest may be missed). database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with similar vectors and distances
delete_vector Delete vector by ID. Args: collection: Vector collection name vector_id: Vector ID to delete database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
add_graph_node Add node to graph. Args: node_id: Node identifier properties: Node properties node_type: Optional node type/label database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with node creation result
add_graph_edge Add edge between nodes in graph. Args: from_node: Source node ID to_node: Target node ID edge_type: Optional edge type/label properties: Edge properties database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with edge creation result
query_graph Query graph with traversal pattern. Args: start_node: Starting node ID pattern: Traversal pattern (e.g., 'neighbors', 'shortest_path') depth: Maximum traversal depth database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with traversal results
delete_graph_node Delete graph node. Args: node_id: Node ID to delete database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
delete_graph_edge Delete graph edge. Args: edge_id: Edge ID to delete database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
set_keyvalue Set key-value pair with optional TTL. Args: key: Key name value: Value to store (any JSON-serializable type) namespace: Namespace for the key (default: "default") ttl: Time-to-live in seconds (optional) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with success status and key details
get_keyvalue Get value by key. Args: key: Key name namespace: Namespace for the key (default: "default") database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with key and value (None if key doesn't exist)
delete_keyvalue Delete key-value pair. Args: key: Key name namespace: Namespace for the key (default: "default") database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
list_keyvalues List all keys in namespace. Args: namespace: Namespace to list keys from (default: "default") database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with keys list and count
publish_stream Publish event to stream. Args: stream_name: Name of the stream event_data: Event payload data database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with publish status and event details
consume_stream Consume events from stream. Events are returned in ascending ingestion order (oldest first), enabling deterministic replay. Use `next_cursor` from the response to page forward without re-scanning events already consumed. Args: stream_name: Name of the stream limit: Maximum number of events to return (default: 10) cursor: Opaque pagination cursor from a previous call's `next_cursor`. Omit (or pass null) to start from the beginning of the stream. database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with events list, count, and next_cursor (null when end-of-stream)
store_spatial Store spatial geometry with attributes. Args: table_name: Name of spatial table geometry: Geometry as WKT string or GeoJSON dict attributes: Optional attributes for the geometry database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with feature_id and geometry details
search_spatial_nearby Search for nearby spatial features within radius. Args: table_name: Name of spatial table longitude: Longitude of search point latitude: Latitude of search point radius_meters: Search radius in meters limit: Maximum number of results (default: 10) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with nearby features and distances
search_spatial_bbox Search spatial features within bounding box. Args: table_name: Name of spatial table min_lon: Minimum longitude min_lat: Minimum latitude max_lon: Maximum longitude max_lat: Maximum latitude limit: Maximum number of results (default: 10) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with features within bounding box
delete_spatial Delete spatial feature by ID. Args: table_name: Name of spatial table feature_id: ID of feature to delete database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
store_timeseries Store time-series data point. Args: metric: Metric name value: Numeric value to store timestamp: ISO timestamp (optional, defaults to now) tags: Optional tags for the data point database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with storage status and point details
query_timeseries Query time-series data over time range with aggregation. Args: metric: Metric name start_time: Start timestamp (ISO format) end_time: End timestamp (ISO format) aggregation: Aggregation function (avg, sum, min, max, count) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with aggregated time-series results
create_analytics_cube Create OLAP analytics cube from fact table. Args: cube_name: Name for the cube fact_table: Source fact table name dimensions: List of dimension column names measures: List of measure column names database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with cube creation status
query_analytics Query analytics cube with dimensional aggregation. Args: cube_name: Name of the cube dimensions: Dimensions to group by (optional) measures: Measures to aggregate (optional) filters: Filter conditions (optional) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with aggregated results
delete_analytics_cube Delete analytics cube. Args: cube_name: Name of the cube to delete database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
put_s3_object Upload object to S3 bucket. Args: bucket: Bucket name key: Object key (path) content: Object content content_type: MIME type (optional) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with upload status and object details
get_s3_object Get object from S3 bucket. Args: bucket: Bucket name key: Object key (path) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with object content and metadata
list_s3_objects List objects in S3 bucket. Args: bucket: Bucket name prefix: Key prefix filter (optional) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with object list and count
delete_s3_object Delete object from S3 bucket. Args: bucket: Bucket name key: Object key (path) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
create_bucket Create an S3 bucket (idempotent — safe to call even if the bucket already exists). Args: bucket: Bucket name (3-63 chars, lowercase letters/digits/hyphens/dots only) region: AWS region hint; falls back to tenant S3 config region database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with created flag. {"created": true} on first creation, {"created": false, "already_existed": true} if bucket already exists.
delete_bucket Delete an S3 bucket. The bucket must be empty before deletion. Args: bucket: Bucket name to delete database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with success flag.
list_buckets List all S3 buckets for the authenticated tenant. Args: database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with buckets list and count. {"buckets": [...], "count": N}
store_temporal Store data with temporal versioning. Args: entity_id: Unique entity identifier data: Data to store timestamp: ISO timestamp (optional, defaults to now) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with version details
query_temporal_at Query temporal data at specific time point. Args: entity_id: Unique entity identifier timestamp: ISO timestamp to query database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with data at specified time
query_temporal_history Query temporal data history over time range. Args: entity_id: Unique entity identifier start_time: Start timestamp (ISO format) end_time: End timestamp (ISO format) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with history of versions
index_fulltext Index document for full-text search. Args: index_name: Name of the FTS index doc_id: Unique document identifier document: Text content to index database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with indexing status
search_fulltext Full-text search with BM25 ranking. Args: index_name: Name of the FTS index query: Search query string limit: Maximum number of results (default: 10) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with ranked search results
delete_fulltext Delete document from full-text index. Args: index_name: Name of the FTS index doc_id: Document identifier to delete database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
store_blob Store binary blob with metadata. Args: content: Blob content (base64 encoded for binary data) filename: Optional filename mime_type: Optional MIME type metadata: Optional metadata dict database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with blob_id and storage details
get_blob Get blob by ID. Args: blob_id: Unique blob identifier database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with blob content and metadata
delete_blob Delete blob by ID. Args: blob_id: Unique blob identifier database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
store_memory Store agent interaction with automatic memory classification. Args: user_input: User's input text llm_response: LLM's response text session_id: Optional session identifier interaction_type: Type (conversation, knowledge_acquisition, task_completion, reasoning) metadata: Optional metadata about the interaction database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with memory_id and classification results
recall_memory Retrieve relevant context for LLM prompt enhancement. Args: query: Query or context for retrieval session_id: Optional session identifier. When provided, recall is scoped to that session — only memories stored under the same session_id are returned. When omitted, recall spans all sessions in this database. context_types: Types of context to include (conversational, semantic, temporal, procedural) max_items: Maximum number of context items to return database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with retrieved memories and context
store_knowledge Store knowledge with automatic classification and indexing. Args: content: Knowledge content to store knowledge_type: Type of knowledge (semantic, procedural, episodic, factual) source: Optional source of the knowledge confidence: Confidence score (0.0-1.0) metadata: Optional metadata database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with knowledge_id and classification results
search_knowledge Query stored knowledge with filtering. Args: query: Query string knowledge_types: Types of knowledge to search min_confidence: Minimum confidence threshold limit: Maximum number of results database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with matching knowledge items
delete_memory Delete specific memory/interaction. Args: memory_id: Unique memory identifier database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
delete_knowledge Remove knowledge entry from knowledge base. Args: knowledge_id: Unique knowledge identifier database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with deletion status
create_iceberg_table Create a new Iceberg table with schema. Args: table_name: Name of the table to create schema: Column definitions as {column_name: data_type} partition_by: Optional list of columns to partition by properties: Optional table properties database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with table_id and table_name
append_iceberg Append records to an Iceberg table. Args: table: Table name to append to data_records: List of row dicts to append database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with snapshot_id and record_count
get_iceberg_snapshot_as_of Time-travel query to retrieve table state at a specific timestamp. Args: table: Table name timestamp: ISO 8601 timestamp to query as of database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with snapshot data at the given timestamp
add_iceberg_column Add a column to an existing Iceberg table (schema evolution). Args: table: Table name column_name: Name of the new column column_type: Data type of the new column required: Whether the column is required (default: False) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with table, column_name, and column_type
expire_iceberg_snapshots Expire old snapshots to reclaim storage. Args: table: Table name retain_last_n: Number of recent snapshots to retain (default: 1) database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with table and retain_last_n
read_iceberg Read rows from an Iceberg table (data, not just snapshot metadata). Supports column projection and time-travel: pass `snapshot_id` OR `as_of_timestamp` (ISO 8601) to read the table state at that point; omit both to read the latest snapshot. Args: table: Table name or table_id. columns: Optional list of columns to project. snapshot_id: Optional snapshot to read from (time-travel). as_of_timestamp: Optional ISO 8601 timestamp to read as of (time-travel). limit: Max rows to return (default 1000, hard max 10000). When the table has more rows, `truncated` is true. database: Optional named database within this tenant. If omitted, falls back to the X-PolyDB-Database header, then the tenant's default, then "polydb". ctx: FastMCP context (injected automatically) Returns: Dict with rows, row_count, truncated, table, snapshot_id.
list_databasesList databases in the tenant's project/cluster. Read-only discovery tool. Returns databases accessible to the current tenant, with the default/current database flagged. To create or delete databases, use the REST API at `/api/databases`. This tool enumerates the named databases available as targets for the ``database`` parameter on every dataplane tool (store_document, query_sql, store_vector, etc.). Pass one of the returned names as ``database='<name>'`` to scope a tool call to that database instead of the tenant default. System databases are not included (the underlying DatabaseRepository queries the user_databases table, never the system catalog). Note: a brand-new tenant whose default DB has not yet been materialized via the REST endpoint will see an empty list here — call POST /api/databases first. Returns: {"databases": [{"name": str, "current": bool, "created_at": str | None}, ...]}
search_schema Discover all available MCP tools with their parameter signatures. Use this tool to explore what operations are available in PolyDB before deciding which tool to call. Returns a compact catalog of all registered tools with required and optional parameters. Args: filter: Optional provider name to filter results (e.g. 'document', 'vector', 'sql') ctx: FastMCP context (injected automatically) Returns: Dict with tools list, total count, and filter_applied
execute_workflowExecute a multi-step workflow in one round-trip. Each step references a tool by name with arguments. Steps execute sequentially. Use $ref(stepId.field) to chain step outputs as inputs to later steps. Args: steps: List of step dicts, each with: - id: unique step identifier - op: tool name (e.g., "store_document") - args: tool arguments dict Named-database routing note (crit-1000): DSL steps inherit the same ``database`` precedence as discrete tool calls — if a step's ``args`` dict includes ``database='<name>'``, that value is consumed by the dispatch boundary (``_instrumented_call_tool``) and overrides the ``X-PolyDB-Database`` header or tenant default for that step only. Resolution order: args.database > X-PolyDB-Database header > tenant default > "polydb".