Data Collection
mostlyright.md
Search and retrieve detailed metadata, schemas, and sample rows from public datasets on Mostly Right.
ENDPOINT 1
https://mostlyright.md/api/mcp
MCP server metadata
- Name
- mostlyright
- Version
- 2.0.0
Mostly Right publishes reproducible datasets as immutable Parquet versions with a public catalog, and can also build new ones. Anonymous, no account: search, fetch, search_datasets, get_dataset, list_tables, get_table_schema, sample_rows, get_download_instructions, get_access_instructions. `search` and `fetch` are the two-tool connector contract — search returns {id, title, url} per dataset and fetch returns that dataset as Markdown. With a workspace credential: query_table, catalog_search and list_connected_datasets. connect_dataset also changes the workspace, so it needs an OAuth connection rather than an mr_use_ key. With a paid subscription, an OAuth connection carrying datasets:build, and the Owner, Admin or Editor role: create_dataset, update_dataset, list_my_datasets, get_my_dataset, register_recipe, start_run, confirm_run, cancel_run, get_run, run_events, list_runs, query_run, run_artifacts, get_artifact_download, write_note, promote_table, get_table and list_source_credentials. A normal read is search_datasets -> get_dataset -> get_table_schema -> sample_rows. If the user needs filtered, sorted or aggregated rows, call query_table; if that refuses, call get_access_instructions and give the user the signup URL and steps it returns rather than reporting a failure. Looking for a source rather than a published dataset? catalog_search reads a sealed snapshot of public sources — one provider, Data.gov, partially swept. Treat a hit as a lead to open and read, and a miss as nothing at all: it is not evidence that no such source exists. A build is: read mostlyright://build-guide and mostlyright://recipe-reference, write a recipe document with a description on every column and units on every numeric column, register_recipe, start_run in sample mode with ceilings, run_events until it is terminal, query_run to check the rows, write_note for each decision. Fix the document and register it again as often as needed; a full run and promote_table happen only when the user says so. A run over the spend threshold comes back held with its projected cost — show the user those numbers and call confirm_run only when they agree. Source credentials are pasted by a person in the web app, at /dashboard/settings/secrets or the dataset's Source credentials panel, and stored under a name. A recipe references the name. Never ask for a secret in chat, never accept one offered there, and never put one in a recipe. Never send SQL to query_table: it takes a closed JSON grammar. query_run does take SQL, over one finished run's own table, where the relation is always named run_table. Never page a whole table through either; get_download_instructions returns the one request that fetches the complete Parquet snapshot. Cite results with the dataset's canonical_url, the table title, the immutable version_id, and the content digest when one is returned.
Known tools 31
get_datasetOne dataset's overview: title, summary, topics, publisher, publication and update dates, canonical page URL, star and table counts, and every table with its id, title, license, immutable version_id, column count, column names and the capabilities the publisher enabled.
Potential side effectslist_tablesThe tables in one dataset, without the full column schemas — the cheap call when you only need table ids and titles.
Inferred read-onlyget_table_schemaOne table's columns (name, type, and any published profile such as null counts, distinct counts or ranges), its immutable version_id, and its capabilities: whether anonymous sampling, keyed querying and Parquet download are enabled.
Inferred read-onlysample_rowsThe publisher's materialized preview of a table — real rows, no account, no query cost.
Inferred read-onlyget_download_instructionsThe exact URL, HTTP method and header for downloading one table's current immutable Parquet snapshot, plus whether the publisher enabled it.
Inferred read-onlyget_access_instructionsWhat this server does anonymously, what needs an account, the two API key classes and what each unlocks, the OAuth option, the signup URL, and the numbered steps from 'no account' to 'querying rows'.
Inferred read-onlysearchSearch every public dataset on Mostly Right and return up to 20 matches as {results: [{id, title, url}]}, where `id` is the dataset slug and `url` its canonical page.
Inferred read-onlyfetchThe full public document for one dataset as Markdown: summary, facts, access instructions, and every table with its columns, types, descriptions and units.
Inferred read-onlycatalog_searchSearch a sealed snapshot of public data sources for feeds that might answer a question — the first move when you need a source and do not already know one.
Inferred read-onlylist_connected_datasetsThe public datasets the authenticated workspace has connected, which is exactly the set query_table and the Parquet download will serve.
Inferred read-onlyconnect_datasetConnects one public dataset to the authenticated workspace so its tables become queryable and downloadable.
Inferred read-onlylist_my_datasetsEvery dataset this workspace owns, most recently updated first — not the public catalog.
Inferred read-onlyget_my_datasetOne workspace dataset in build terms: name, description, status, each table with its build state, promotion state and live version, the latest run, and any run held waiting for a spend confirmation.
Inferred read-onlyregister_recipeRegisters the JSON recipe that says where the data comes from, how it is shaped and what must be true of it.
Inferred read-onlyget_runOne run's current state: status, mode, the clamps it ran under, rows delivered and whether a clamp truncated them, how many pages a many-page source reached, the table and table version it sealed, the failure code, detail and stage when it failed, and the version number a confirm or cancel should send.
Potential side effectsrun_eventsOpens the run's event stream, collects up to max_events, and returns as soon as the run reaches a terminal event or wait_seconds elapses — whichever comes first.
Potential side effectslist_runsThe workspace's own runs, newest first, optionally narrowed to one dataset, one status or one mode.
Inferred read-onlyquery_runRuns one read-only SQL statement over the parquet a run sealed, and waits for the answer.
Inferred read-onlyrun_artifactsThe files a run sealed: the parquet, the column profile, the receipt, the preview.
Inferred read-onlywrite_noteAppends one cell to a run's or a dataset's decision record — why a source was chosen, what a check found, what you changed and why.
Inferred read-onlyget_tableOne table as the builder sees it: whether it is promoted, which version is live, the refresh schedule and what the platform has learned about its rhythm, when it last refreshed and why it last failed.
Inferred read-onlylist_source_credentialsThe NAMES of the API keys and passwords this workspace has stored for its sources, with their status and when they were added.
Inferred read-onlyCONNECT WITH APPROVAL
Client installation
Review this server and its permissions before adding it. Secret placeholders must be set locally.
Codex
~/.codex/config.toml
[mcp_servers.mostlyright]
url = "https://mostlyright.md/api/mcp"
enabled = trueClaude Code
.mcp.json
{
"mcpServers": {
"mostlyright": {
"type": "http",
"url": "https://mostlyright.md/api/mcp"
}
}
}Claude Desktop
Settings → Connectors → Add custom connector
Name: mostlyright
Remote MCP URL: https://mostlyright.md/api/mcpAdd this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.
Cursor
.cursor/mcp.json
{
"mcpServers": {
"mostlyright": {
"url": "https://mostlyright.md/api/mcp"
}
}
}Visual Studio Code
.vscode/mcp.json
Add to Visual Studio Code{
"servers": {
"mostlyright": {
"type": "http",
"url": "https://mostlyright.md/api/mcp"
}
}
}Generic MCP
Client-specific MCP configuration
{
"name": "mostlyright",
"transport": "streamable-http",
"url": "https://mostlyright.md/api/mcp"
}MCP Inspector
Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.
TRUST AND VERIFICATION EVIDENCE
Loading Trust v2 evidence…
Checking the associated registrable domain. The BuiltWith key remains server-side.
Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.