datacharter
This server provides read-only, governed access to data sources via SQL and certified metrics.
List data sources – see configured sources and their types.
List tables – enumerate queryable tables with fully-qualified relation names.
Describe table – inspect columns and types for a given relation.
Run read-only SQL queries – execute SELECT-style queries against the data, with guidance to use LIMIT and fully-qualified names.
List certified metrics – view available governed metrics, their definitions, dimensions, and time-grain support.
Run certified metrics – query governed metrics by name, optionally grouped by dimensions and/or time grain (day, week, month, quarter, year).
Allows querying DuckDB databases as a data source via read-only SQL.
Integrates with GitHub Actions to run data quality tests and schema drift checks in CI.
Allows querying MySQL databases as a data source via read-only SQL.
Enables local agent queries using Ollama-hosted models, with full governance and no data leaving the machine.
Supports any OpenAI-compatible agent for natural language querying of connected data sources.
Allows querying Snowflake data warehouses as a data source via read-only SQL.
Allows querying SQLite databases as a data source via read-only SQL.
DataCharter
Query all your data locally. Hand agents exactly the columns you choose.
A local SQL workspace over files and databases, federated by DuckDB and
governed by charter.yaml. Agents get read-only, PII-masked access to what
the contract grants. Apache-2.0. No paid edition.
datacharter.dev · Docs · Desktop · GovBench · Deploy
Two ways in. Same kernel.
Laptop
Drop a file. Query it. Chart it. An agent sees ••• where PII lives.
uvx datacharter serve # demo workspace, http://127.0.0.1:8321
# or: datacharter init --from && datacharter serve
# or: datacharter init --template life && datacharter serve --localNo terminal: desktop app (beta). macOS (Apple Silicon) or Windows. Unsigned until Apple secrets exist.
brew install datacharter/tap/datacharter # macOS
pip install datacharter # Python 3.11+The workspace is a directory: charter.yaml, queries/, guides/. Commit it.
Secrets stay out. Optional agent: SpaceXAI, Claude Code, Ollama (--local),
or any OpenAI-compatible endpoint.
Related MCP server: orchard-data-mcp
Company
The same binary, on a shared MCP endpoint. Identities live in git, not on our servers. No rows leave your infrastructure.
datacharter mcp --http --host 0.0.0.0 # OAuth env required off loopback
helm install datacharter ./chart \
--set oauth.issuer=... --set oauth.audience=... --set oauth.jwksUri=...
datacharter govbench --json # cite corpus govbench-v1MCP Streamable HTTP (
POST /mcp) with optional OAuth 2.1principals:andgrants:incharter.yaml(default-deny on HTTP)Helm chart and OCI image
Hash-chained audit plus SIEM JSON/OTLP
GovBench: frozen 28-attack corpus, grade A-F
Wrap someone else's MCP server: datacharter mcp --guard "npx -y some-mcp-server".
What the contract enforces
PII default-deny, row filters, plain-English policies (aggregates only),
canaries, a flight recorder, datacharter redteam, access diff on PRs.
CLI reference: docs/cli.md. Security: docs/security.md.
Status: pre-release. V1 in development.
Privacy
Runs on your machine or in your cluster. No telemetry. No DataCharter-operated data plane. Privacy Policy.
License
Available Tools
6 toolsdescribe_tableDescribe tableARead-onlyIdempotent
Show columns and types for one relation (e.g. 'crm.customers').
Inputs: relation (required).
list_metricsList certified metricsARead-onlyIdempotent
List certified metrics — name, what each computes, its dimensions, and whether it supports a time grain. Prefer query_metric over writing SQL when a metric matches the question.
Inputs: no named parameters.
list_sourcesList data sourcesARead-onlyIdempotent
List configured data sources with their types.
Inputs: no named parameters.
list_tablesList tablesARead-onlyIdempotent
List all queryable tables with their fully-qualified relation names.
Inputs: no named parameters.
queryRun read-only SQL queryARead-onlyIdempotent
Run a read-only SQL query and return rows. Use fully-qualified relation names. Prefer LIMIT for exploration.
Inputs: sql (required).
query_metricRun a certified metricARead-onlyIdempotent
Run a certified metric by name and return its governed result. Optionally group by dimensions (by) and/or a time grain.
Inputs: name (required), by, grain.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 6 tools
Each tool targets a distinct action: enumerating sources, tables, or metrics; describing schema; running raw SQL; or running a governed metric. The only near-overlap, query vs query_metric, is explicitly disambiguated by naming and by the guidance in list_metrics.
Tool names follow a predictable verb_noun pattern: list_* for enumeration, describe_table for schema inspection, and query/query_metric for data retrieval. Naming is uniform and intuitive.
Six tools is a well-scoped number for a data discovery and querying server. Each tool covers a necessary step in the workflow—source discovery, table discovery, schema inspection, raw query, metric discovery, and metric query.
The tool set provides a complete read-only analytics flow: discover sources and tables, inspect schemas, run SQL, list certified metrics, and query those metrics. There are no obvious missing operations for its stated purpose.
Maintenance
Related MCP Connectors
Query your warehouse or a CSV with Claude/ChatGPT over MCP, governed by table-level ACL + audit.
Query your org's data in natural language — read-only MCP access to SQL, NoSQL, files & warehouses.
Query, join, profile, clean and convert CSV/JSON/Parquet with server-side DuckDB over MCP.
Hosted MotherDuck MCP server: explore data, run DuckDB SQL, and build Dives and Flights.
Related MCP Servers
- AlicenseNot gradedqualityFmaintenanceRead-only MCP server for SQL databases (SQL Server, Postgres, SQLite) with multi-server support and three-layer safety using AST validation and linting.MIT
- AlicenseAqualityCmaintenancePay-per-call MCP server for market data with USDC micro-payments, featuring unique ML market-regime forecasts, news catalyst radar, and a live-verified algorithmic trading track record.1268 npmMIT
- AlicenseAqualityDmaintenanceA read-only DuckDB MCP server offering context-efficient analytics tools (list_datasets, describe_table, profile_column, explain, query) with a semantic layer for business rules, security guards, and disclosed truncation to help LLMs produce correct answers while minimizing token usage.5MIT
- AlicenseNot gradedqualityBmaintenanceA safe, read-only MCP server for local development databases (PostgreSQL and SQLite) that provides guarded introspection tools like schema listing, sample rows, and explain plans with column masking and query budgets.MIT