fitvete-food-cli

Command-line interface for the FitVete Food & Nutrition API. Zero dependencies (Node 18+, built-in fetch). Prints clean JSON to stdout, so it pipes into jq and is trivial for AI agents to parse. Use --pretty for humans.

Install

# install globally from npm
npm install -g fitvete-food-cli
fitvete-food search-foods chicken

# or run without installing
npx fitvete-food-cli search-foods chicken

Authenticate

Get a free key at https://fitvete.com/api/, then:

export FITVETE_API_KEY=fv_live_your_key
# or pass --key on any command

Commands

| Command | Description | | --- | --- | | search-foods <query...> [--number N] | Search foods; calories, macros, confidence. | | compute <ingredient...> | Full nutrient breakdown for an ingredient list. | | recipes [query...] [filters] | Search recipes by name + filters (--diet --cuisine --type --intolerances --include-ingredients --exclude-ingredients --max-ready-time --min-calories --max-calories --min-protein --sort --offset --number). | | recipe-random [--number N] [--tags ...] | Random recipes. | | recipe-autocomplete <query...> [--number N] | Recipe title suggestions. | | recipes-by-nutrients [--min-protein N --max-calories N ...] | Recipes matching nutrient ranges. | | recipe-similar <id> [--number N] | Recipes similar to an id. | | recipe-diets, recipe-meal-types, recipe-cuisines | Supported filter values. | | recipe <id> [--servings N] | Full recipe detail (ingredients, steps, nutrition); optional servings scaler. | | drinks <query...> [--number N] | Search drinks/cocktails. | | score <name...>\|--barcode CODE [--goals ... --diet ... --allergies ... --avoid ... --explain] | Personalized 0–100 health score + grade. | | analyze-ingredients <text...> | Additives, allergens, NOVA level, diet flags. | | parse <text...> | Free-text meal → structured items + nutrition (NLP). | | alternatives <name...>\|--barcode CODE [--number N] | Healthier swaps, ranked. | | foods-autocomplete <query...> [--number N] | Food name typeahead. | | barcode <code> | Look up a packaged food by UPC/EAN barcode. | | photo <image-path> | Estimate meal nutrition from a JPEG, PNG, or WebP image. | | label <image-path> | Parse a Nutrition Facts label from a JPEG, PNG, or WebP image. | | tools | Print a JSON tool manifest for AI agents (function-calling). | | help, version | Help / version. |

Options

• --key <key> — API key (overrides FITVETE_API_KEY).
• --number, -n <N> — result count for search commands (1–25).
• --base <url> — API base URL (overrides FITVETE_API_BASE).
• --pretty — human-readable table instead of JSON.
• --json — force JSON (default).

Exit codes

0 ok · 1 request/HTTP error · 2 usage error · 3 auth or rate-limit.

Examples

fitvete-food search-foods "greek yogurt" --number 3
fitvete-food compute "1 cup rice" "2 eggs" | jq '.calories'
fitvete-food recipe 715538 | jq '.title'
fitvete-food drinks margarita --pretty
fitvete-food barcode 012345678905
fitvete-food photo ./meal.jpg
fitvete-food label ./nutrition-label.jpg | jq '.nutrients[] | select(.name=="Sodium")'

Using it from an AI agent

The CLI is designed to be driven by autonomous agents:

• Output is pure JSON — no prompts, spinners, or colour codes.
• Failures return a non-zero exit code and a JSON error body.
• fitvete-food tools prints a function-calling manifest describing every command and its parameters, so a model can discover the available functions at runtime.

fitvete-food tools     # → { "tools": [ { "name": "fitvete_search_foods", "input_schema": {...} }, ... ] }

As an MCP tool

Expose the CLI to any MCP client (Claude Desktop, Cursor, …) by running it from a shell tool, or wrap each command as an MCP tool. Minimal stdio wrapper sketch:

// mcp-server.js (pseudocode) — map each manifest tool to a CLI call
import { execFile } from "node:child_process";
const run = (cmd, args) => new Promise((res, rej) =>
  execFile("fitvete-food", [cmd, ...args], (e, out) => (e ? rej(e) : res(JSON.parse(out)))));
// tool "fitvete_search_foods" -> run("search-foods", [query, "--number", String(n)])

The same key (FITVETE_API_KEY) and JSON contract apply.

License

MIT