@cyanheads/bls-labor-mcp-server

Public Hosted Server: https://bls-labor.caseyjhand.com/mcp

Tools

Seven tools in two groups — four for BLS data access (survey discovery, SeriesID resolution, history, current values) and three for optional DataCanvas SQL analysis of large result sets:

bls_list_surveys

List available BLS survey programs and their metadata.

• Covers all major BLS programs: CPS, CES, CPI, PPI, JOLTS, LAUS, OEWS, ECEC, and others

• Optional category filter (prices, employment, wages, productivity, injuries, time_use)

• Returns survey codes, descriptions, and calculation-support flags (allowsNetChange, allowsPercentChange, hasAnnualAverages)

• Backed by the live BLS surveys API with monthly caching; does not consume daily API quota

bls_search_series

The entry point for most BLS workflows. Resolves human concepts to BLS SeriesIDs.

• BLS identifiers like LNS14000000 and CES0000000001 encode survey + area + item + seasonal flag in opaque positional codes — this tool decodes them

• Free-text and keyword search against the full BLS series catalog

• Filter by survey code, geographic area (state name, MSA, or FIPS), and seasonal adjustment flag

• Returns decoded series components (survey, area, item, seasonal flag) alongside the plain-language name

• Operates entirely offline against LABSTAT flat files bundled at startup — no API quota consumed

• Use before bls_get_series or bls_get_latest when you have a concept but not a SeriesID

bls_get_series

Fetch historical time-series data for one or more BLS series.

• Batch fetch up to 50 series per request (counts as one of the 500 daily API queries)

• Optional start_year / end_year window (BLS caps history at 20 years per request)

• Optional calculations: true for BLS-server-side net change and percent change — a survey returns whichever it supports (CPI/PPI return percent change only); check bls_list_surveys for per-survey support

• Optional annual_average: true adds each year's annual-average row (period M13, Q05 or S03) — the mean of that year's real periods, not an additional one. Off by default, so observations are safe to sum or average as returned; enrichment.annualAverageRows reports how many were added

• Returns observations with series metadata (title, area, item, seasonality)

• Spills to a DataCanvas dataframe when the observation count exceeds the inline context budget — response includes a dataset.name handle for SQL via bls_dataframe_query. Requires CANVAS_PROVIDER_TYPE=duckdb.

bls_get_latest

Get the current value for one or more BLS series.

• Issues one GET per SeriesID (no batch-latest endpoint exists in BLS v2) — each counts as one of the 500 daily API queries

• Recommended limit: ≤10 series per call; accepts up to 50

• For "current value" across many series, bls_get_series with a narrow year window is more quota-efficient (one API query regardless of series count)

• Partial success reporting — failed series are returned in a separate failed[] array alongside successful results

bls_dataframe_describe

Inspect canvas dataframes registered by bls_get_series.

• Lists all active dataframes for the current tenant: table name, source tool, query params, row count, column schema, TTL

• Optionally describe a single dataframe by name

• Lazy-sweeps expired entries before responding

• Use before writing SQL to confirm column names

bls_dataframe_query

Run SQL against canvas dataframes registered by bls_get_series.

• Read-only: writes, DDL, DROP, COPY, PRAGMA, ATTACH, and external-file table functions are rejected

• Supports JOINs, aggregates, window functions, and CTEs

• Optional register_as persists the query result as a new named dataframe with a fresh TTL — useful for chaining analyses without re-consuming BLS API quota

• Inline row cap: 1,000 rows by default (max 10,000); full results live on-canvas when register_as is set

• Zero BLS API quota consumed

bls_dataframe_drop

Drop a canvas dataframe by name. Idempotent — returns dropped: false when nothing matched.

• Use to free canvas resources ahead of the per-table TTL when an analysis is complete

• Must be explicitly enabled via BLS_DATAFRAME_DROP_ENABLED=true (TTL handles cleanup by default)

Related MCP server: mcp-fred

Features

Built on @cyanheads/mcp-ts-core:

• Declarative tool definitions — single file per tool, framework handles registration and validation

• Unified error handling across all tools

• Pluggable auth (none, jwt, oauth)

• Swappable storage backends: in-memory, filesystem, Supabase, Cloudflare KV/R2/D1

• Structured logging with optional OpenTelemetry tracing

• STDIO and Streamable HTTP transports

BLS-specific:

• BLS API v2 integration with retry/backoff and daily quota tracking

• Offline series catalog search against LABSTAT flat files — zero API quota for discovery

• Typed error contracts for BLS-specific failure modes: quota exhaustion, locked database, calculations not supported

• Period-over-period calculations via BLS server-side flag (consistent with BLS published numbers)

• DataCanvas spillover (DuckDB) for large multi-series result sets — SQL access without re-querying the API

• Optional local observation mirror — sync LABSTAT bulk data into an embedded SQLite store to serve bls_get_series / bls_get_latest without the 500/day API cap (opt-in, off by default)

• On-disk SQLite catalog index — the series catalog is parsed into an FTS5 SQLite store, queried on demand (not held in memory) and persisted across restarts; the OES/OEWS wage survey (~6M series) is opt-in via BLS_CATALOG_INCLUDE_OES

Getting started

Add the following to your MCP client configuration file. A free BLS API key unlocks 500 queries/day — register at bls.gov/developers. The server works without a key at 25 req/day.

{
  "mcpServers": {
    "bls-labor-mcp-server": {
      "type": "stdio",
      "command": "bunx",
      "args": ["@cyanheads/bls-labor-mcp-server@latest"],
      "env": {
        "MCP_TRANSPORT_TYPE": "stdio",
        "MCP_LOG_LEVEL": "info",
        "BLS_API_KEY": "your-key-here"
      }
    }
  }
}

Or with npx (no Bun required):

{
  "mcpServers": {
    "bls-labor-mcp-server": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "@cyanheads/bls-labor-mcp-server@latest"],
      "env": {
        "MCP_TRANSPORT_TYPE": "stdio",
        "MCP_LOG_LEVEL": "info",
        "BLS_API_KEY": "your-key-here"
      }
    }
  }
}

Or with Docker:

{
  "mcpServers": {
    "bls-labor-mcp-server": {
      "type": "stdio",
      "command": "docker",
      "args": ["run", "-i", "--rm", "-e", "MCP_TRANSPORT_TYPE=stdio", "-e", "BLS_API_KEY=your-key-here", "ghcr.io/cyanheads/bls-labor-mcp-server:latest"]
    }
  }
}

For Streamable HTTP, set the transport and start the server:

MCP_TRANSPORT_TYPE=http MCP_HTTP_PORT=3010 BLS_API_KEY=... bun run start:http
# Server listens at http://localhost:3010/mcp

Prerequisites

• Bun v1.3.2 or higher (or Node.js v24+).

• A free BLS API v2 key — register at bls.gov/developers. Grants 500 queries/day; the server also works without a key at 25 req/day.

Installation

1. Clone the repository:

git clone https://github.com/cyanheads/bls-labor-mcp-server.git

2. Navigate into the directory:

cd bls-mcp-server

3. Install dependencies:

bun install

4. Configure environment:

cp .env.example .env
# edit .env and set BLS_API_KEY

Configuration

All configuration is validated at startup via Zod schemas in src/config/server-config.ts.

See .env.example for the full list of optional overrides.

Observation mirror (optional)

For high-volume workloads, an opt-in local mirror serves bls_get_series / bls_get_latest from an embedded SQLite store instead of the BLS API — eliminating the 500/day quota cap. It is off by default. To enable:

1. Set BLS_OBSERVATIONS_MIRROR_ENABLED=true (and review the BLS_OBSERVATIONS_MIRROR_* vars in .env.example).

2. Run the one-time bootstrap out-of-band — it downloads the full LABSTAT observation set and can take a while:

   node dist/services/bls-observations/subprocess.js --init

Until the bootstrap completes, requests fall back to the live API (unless BLS_OBSERVATIONS_MIRROR_FALLBACK_LIVE=false). On HTTP transport, an incremental refresh runs on the BLS_OBSERVATIONS_MIRROR_REFRESH_CRON schedule. In containers, mount a persistent volume at BLS_OBSERVATIONS_MIRROR_PATH.

Running the server

Local development

• Build and run:

  # One-time build
  bun run rebuild
  
  # Run the built server
  bun run start:stdio
  # or
  bun run start:http

• Run checks and tests:

  bun run devcheck   # Lint, format, typecheck, security
  bun run test       # Vitest test suite
  bun run lint:mcp   # Validate MCP definitions against spec

Docker

docker build -t bls-labor-mcp-server .
docker run --rm -e BLS_API_KEY=your-key -e MCP_TRANSPORT_TYPE=http -p 3010:3010 bls-labor-mcp-server

The Dockerfile defaults to HTTP transport, stateless session mode, and logs to /var/log/bls-labor-mcp-server. OpenTelemetry peer dependencies are installed by default — build with --build-arg OTEL_ENABLED=false to omit them.

Project structure

Development guide

See CLAUDE.md for development guidelines and architectural rules. The short version:

• Handlers throw, framework catches — no try/catch in tool logic

• Use ctx.log for request-scoped logging, ctx.state for tenant-scoped storage

• bls_search_series is the anchor tool — design workflows to call it before the API tools

• Wrap BLS API calls: validate raw → normalize to domain type → return output schema; never fabricate missing fields

Contributing

Issues and pull requests are welcome. Run checks and tests before submitting:

bun run devcheck
bun run test

License

Apache-2.0 — see LICENSE for details.