Skip to main content
Glama
eait-fit

fooddb

Official
by eait-fit
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

fooddb

A global food database: generic nutrition and branded products by barcode, kept current and served as a REST API, an MCP server and a CLI.

Status: early prototype. Fetchers for USDA FDC and Open Food Facts, async jobs on pq, and a REST API.

Run locally

Needs Docker and uv; ./dev install installs what is missing.

./dev up --all       # shared Postgres, this worktree's databases, API + pq worker
./dev fetch all      # queue USDA FDC (Foundation, SR Legacy) and 7 Open Food Facts deltas
./dev jobs           # row counts, fetch runs, queue state
./dev status         # this worktree;  ./dev ls  for every worktree
./dev test           # unit tests + the database suite against this worktree's __test database
./dev down

./dev with no arguments lists every command. Each git worktree gets its own slot: the API port (9640 + 10 × slot) and its own dev and __test databases on one shared Postgres (port 5440). All of these are derived into .env.worktree by scripts/dev_env.py.

curl "$(./dev url)/v1/foods?q=hummus"                         # search products
curl "$(./dev url)/v1/products/06297001181102?include=off"    # by barcode, with the OFF layer
curl "$(./dev url)/v1/records/fdc:168421"                     # the product a source record belongs to
curl "$(./dev url)/v1/foods/1?snapshot=2026-10-04"            # pinned to a day's snapshot
curl "$(./dev url)/healthz"                                   # freshness; 503 when stale

How data flows: fetchers write append-only observations per source record; values that fail a check wait for review; Splink matches records into products; the nightly snapshot freezes each product's resolved values (most trusted source per field), and the API serves that snapshot.

Fetcher

Source

Licence

Schedule

fdc foundation, sr_legacy

USDA FoodData Central bulk JSON

CC0

weekly check (USDA releases twice a year)

off

Open Food Facts daily delta files

ODbL, off layer

every 6 hours

off-dump

Open Food Facts full dump (~13 GB, streamed)

ODbL, off layer

once, then deltas

match

Splink product matching

–

after every fetch that added data

snapshot

Nightly snapshot of resolved values

–

02:30 daily

uv run pytest runs the unit tests alone; the database suite is skipped without ./dev test.

Related MCP server: Open Food Facts MCP Server

Contributing

See CONTRIBUTING.md. Report vulnerabilities privately: SECURITY.md.

Licence

The code is under AGPL-3.0. The data is licensed separately: the core data under fooddb's own terms, and the Open Food Facts layer under ODbL.

Relation to eait

eait (eait.fit) is the first customer and the first data source:

  • Customer. eait keeps a read-only local copy of the catalog (food_ref / off_product) and refreshes it from the nightly snapshot. It never calls this service live, so eait keeps logging meals when fooddb is down.

  • Data source. When a barcode scan in eait misses or hits a stale row, eait sends the label photo here as an observation. No user id crosses the boundary.

This server cannot be deployed

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Related MCP Connectors

Related MCP Servers

  • Open Food Facts MCP Server

    caleb-conner
    A
    license
    A
    quality
    D
    maintenance
    Enables AI assistants to access the Open Food Facts database to query detailed food product information, nutritional data, and environmental scores. Supports product lookup by barcode, smart search with filtering, nutritional analysis, product comparison, and dietary recommendations to help users make informed food choices.
    5
    1
    MIT

  • USDA Nutrition Database MCP

    HemantaPatil
    F
    license
    Not graded
    quality
    D
    maintenance
    Provides intelligent access to the USDA nutrition database through AI assistants, enabling users to search foods, compare nutritional content, find foods high in specific nutrients, and query authoritative nutrition data across 7,146+ food items through natural language.
    1
    -