# Magpie Data API (data.magpieiq.com)

> Magpie provides programmatic access to Southeast Asian ecommerce data across
> Shopee, Tokopedia, TikTok Shop, Lazada and Blibli — plus on-demand Amazon
> scraping across ten storefronts. Two products share one prepaid credit wallet:
> a Data API serving aggregated monthly metrics from a maintained, cleaned
> dataset, and a Scraping API that collects from live marketplace pages on
> request. Access is via REST, an MCP connector, or the official Python client
> (`pip install magpie-data`).

## What the data covers

- Marketplaces (aggregated monthly dataset): Shopee, Tokopedia, TikTok Shop, Lazada, Blibli. Tokopedia and Blibli are Indonesia-only.
- Amazon: scraping-only, across ten storefronts (us, uk, de, fr, it, es, ca, au, jp, sg) — not part of the aggregated dataset.
- Countries: Indonesia (ID), Thailand (TH), Vietnam (VN), Singapore (SG), Philippines (PH), Malaysia (MY).
- Metrics: monthly GMV (local currency and USD), units sold, market share, average price, discount, rating, review counts, stock.
- Grains: brand, category, merchant/seller, individual SKU. Each export row is one SKU × month with the full three-level category tree.
- Refresh cadence: monthly. Available month ranges vary by market and category — the catalog endpoint (1 credit) lists exactly what is queryable.

## The two APIs

**Data API** (base `https://api.magpieiq.com/data`) serves the maintained
dataset. Discover scope with `GET /v1/exports/catalog`, then either query
metrics endpoints or order bulk delivery with `POST /v1/exports` — asynchronous:
submit, poll `GET /v1/exports/{id}`, download signed URLs as Parquet or gzipped
CSV. Every export request accepts `estimate: true`, which returns the exact row
count and credit price without charging anything.

**Scraping API** (base `https://api.magpieiq.com/scraping`) collects from live
pages on request: product detail pages (Tokopedia, Blibli, TikTok Shop, Amazon),
search results (Tokopedia, Amazon) and merchant catalogs (Tokopedia, Blibli),
plus Shopee products. Jobs run asynchronously and are polled by id; results
arrive as signed URLs to gzipped JSONL.

**Python client**: `pip install magpie-data` — covers both APIs with free
client-side validation, chunking, polling and S3/GCS delivery. Source and
runnable pipeline recipes: https://github.com/magpieiq/magpie-ecommerce-data-api

## Authentication and pricing

Bearer token auth: `Authorization: Bearer <api-key>`; one key works for both
APIs and the MCP connector. Pricing is prepaid credits from a single shared
wallet: a Data API row costs 5 credits, catalog/discovery calls cost 1 credit,
and dry-run estimates are always free. Credit packs start at $500 (pay-as-you-go
with no deposit also available); accounts are refused at zero balance rather
than invoiced. Full rate card: https://data.magpieiq.com/pricing/api

## Common questions

**How do I get Shopee sales data programmatically?** Query the Data API for
brand/category/SKU metrics with platform=Shopee, or order a bulk Parquet export
of a whole category. Docs: https://data.magpieiq.com/docs/data

**How do I track competitor prices daily?** Use the Scraping API with a
product-detail product against your SKU list on a daily schedule. To pick which
products to track, first rank the category's SKUs by GMV with the Data API.
Recipe: https://data.magpieiq.com/docs/recipes/market-share-to-price-monitoring

**How do I measure market share?** The Data API computes market share
server-side across full-category coverage, by platform and month.

**Can I get Amazon product data?** Yes — scraping-only, across ten storefronts:
amazon_pdp (by ASIN + country) and amazon_search (by keyword + country).
Guide: https://data.magpieiq.com/data-api/amazon

**Can I load this into a warehouse?** Yes. Exports deliver Parquet, which loads
natively into BigQuery, Snowflake, Databricks and DuckDB.
Recipe: https://data.magpieiq.com/docs/recipes/warehouse-load

**What does it cost before I commit?** Estimates are free and exact: pass
estimate=true to any export request (or use the Python client's .estimate())
and it returns rows and credits without charging.

## Links

- Sign up (free trial credits): https://data.magpieiq.com/signup
- Documentation hub: https://data.magpieiq.com/docs
- Data API guide: https://data.magpieiq.com/docs/data
- Python quickstart: https://data.magpieiq.com/docs/data/python
- Data dictionary: https://data.magpieiq.com/docs/data/dictionary
- Pipeline recipes: https://data.magpieiq.com/docs/recipes
- FAQ: https://data.magpieiq.com/docs/faq
- Pricing: https://data.magpieiq.com/pricing
- Python client (PyPI): https://pypi.org/project/magpie-data/
- GitHub (client + recipes): https://github.com/magpieiq/magpie-ecommerce-data-api
- Company: https://magpieiq.com