This comparison is for KRW traders, quant developers, and analysts deciding whether to build Korean macro data straight from KOSIS, the Korean Statistical Information Service, or to use FXMacroData as the data layer for USD/KRW research. KOSIS is the official national statistics portal run by Korea's Ministry of Data and Statistics (Statistics Korea until October 2025), and its OpenAPI is free once you register for a key. FXMacroData is a paid FX macro API that reads the same official Korean releases and packages them for trading workflows.
The core finding: KOSIS wins on depth, authority, and cost. It exposes hundreds of thousands of official tables, statistical metadata, and an SDMX option at no charge. FXMacroData wins when the job is a currency decision rather than a statistics project: one schema across 22 currencies, a second-level announcement_datetime on every release row, a release calendar, pair dashboards, and MCP access for AI assistants. Many KRW desks will use both: KOSIS for the deep table, FXMacroData for the release-time record they trade on.
Decision snapshot
Choose KOSIS OpenAPI when
You need the full official Korean table: every CPI sub-item, regional splits, survey metadata, or a long research history, and you are comfortable working with Korean-language documentation.
Choose FXMacroData when
You trade KRW against other currencies and need release-timed headline series, a calendar, pair context, and the same JSON shape for Korea, the US, Japan, and 19 other currencies.
Quick answer
KOSIS is free and authoritative for Korean statistics. FXMacroData costs $50/month (Individual) and is built for KRW trading workflows that need publication timestamps and cross-currency consistency.
Takeaway: the split is source of record versus trading workflow layer, not good data versus bad data.
Pricing And Access
KOSIS is not a commercial vendor, so the price comparison is lopsided by design. The KOSIS OpenAPI terms state that information is provided free of charge in principle, with fees only if a user asks for delivery outside the standard service. Access needs a KOSIS membership and an approved usage application, which issues the API key passed as the apiKey parameter on each request. The Bank of Korea runs a separate free, key-based API for its own statistics database (ECOS), covered in our Bank of Korea and KRW analysis.
FXMacroData publishes self-serve pricing. The Always Free tier needs no key, covers USD data only, allows 100 requests a day, and delays USD announcements by 15 minutes. Korean data starts with the Individual plan at $50/month or $500/year, with a 14-day trial, 1,000,000 requests a month, and bursts of 300 a minute. Business is $250/month or $2,500/year for companies and teams of up to 10 people, with the same monthly allowance and priority support. Enterprise is by custom agreement.
KOSIS OpenAPI
Free, key required*
Membership plus usage application. Up to 200 calls a minute and 40,000 cells per request on the standard services.*
FXMacroData Individual
$50/month
KRW plus 21 other currencies, 14-day trial, 1,000,000 requests a month, dashboards, exports, and MCP.
FXMacroData Always Free
$0, no key
USD data only, 100 requests a day, announcements delayed 15 minutes. Useful for testing the schema before paying.
Takeaway: if budget is the only constraint, KOSIS is the answer. The $50 buys the engineering you would otherwise do yourself: normalisation, release timing, and cross-currency consistency.
Side-By-Side Comparison
| Attribute | FXMacroData | KOSIS OpenAPI |
|---|---|---|
| Core job | FX macro API: release-timed indicator rows, calendar, pair dashboards, exports, MCP. | Official national statistics portal: tables, metadata, survey descriptions, key indicators. |
| Entry pricing | Individual $50/month or $500/year; Business $250/month for up to 10 people. | Free; fees only for non-standard delivery requests.* |
| Free tier | Always Free: USD only, no key, 100 requests/day, 15-minute delay on USD announcements. KRW needs a paid plan. | The whole service is free after membership and key approval.* |
| Coverage | 22 currencies, 700+ series. KRW covers CPI, PPI, GDP, policy rate, trade and current account, labour, retail sales, confidence, housing, credit, and government bond yields. | Korea only, but very deep: the portal lists 290,552 statistical tables (count dated June 2023), 1,529 survey descriptions, and 1,469 key indicators.* |
| Release timing | Second-level announcement_datetime on every row; target of one second from publication, measured publicly. |
Data rows carry the period and a last-changed date (LST_CHN_DE), not a publication timestamp.* |
| API format | REST JSON with OpenAPI, header auth (X-API-Key), SSE streaming, GraphQL, Python/JS SDKs. |
REST over HTTPS with key as a query parameter; JSON, SDMX, XML, or XLS depending on the service.* |
| Rate limits | 1,000,000 requests/month; bursts of 300/minute and 5,000/hour; 20 concurrent. | 200 calls per minute; 40,000 cells per request (200,000 for bulk XLS).* |
| History | Multi-year history per series; each response reports its earliest_available_date. |
Full published history of each table, back to the start of the series. |
| Language | English docs and field names. | OpenAPI portal and guide in Korean; responses include English labels (ITM_NM_ENG, UNIT_NM_ENG).* |
| AI / MCP | Official MCP server for Claude, ChatGPT, Cursor, and other clients. | No MCP server listed on the OpenAPI portal. |
| Licence for programmatic / commercial use | Plans cover internal use; showing data to outside users needs the self-serve Commercial Redistribution add-on ($10/month per block of up to 100 users). | Commercial use allowed, with source attribution required; reselling the raw data without your own service platform is prohibited.* |
| Best fit | KRW and multi-currency traders, event-study builders, dashboards, alerting, AI agents. | Korea economists, academic research, sub-item and regional analysis, official citation. |
* Competitor terms and limits retrieved from the KOSIS OpenAPI service-introduction, developer-guide, and terms-of-use pages in October 2026.
What KOSIS Covers And How It Works
KOSIS is Korea's national statistics database, bringing together official statistics from the Ministry of Data and Statistics and other producing agencies, the Bank of Korea included. The OpenAPI offers seven services: statistics lists, statistical data, bulk statistical data, survey descriptions, table descriptions, integrated search, and key indicators. All of them use REST over HTTPS. The portal ended plain-HTTP access and announced per-minute call limits in a February 2026 notice, then followed up with a July 2026 notice that added statistical symbols to the data services.
Requests are built around table IDs and classification codes. A data call names the organisation (orgId), the table (tblId), up to eight classification levels (objL1 to objL8), the item (itmId), the period type (prdSe), and either a date range or a count of recent periods. Alternatively, you can register a table in the portal and call it by a user statistics ID. The developer guide is in Korean, but responses include English labels for items, classifications, and units, so an English-speaking developer can get started with some help from a translator.
For a KRW trader the important tables are few: headline and core consumer prices, producer prices, the labour force survey, industrial activity, and retail sales. The challenge is locating the right table and item codes, then keeping them current when a table is rebased or reorganised. The portal's notice board regularly announces table changes, and the terms make watching those notices the user's responsibility.
Korean Inflation From The Official Release
Korean consumer prices are released by the Ministry of Data and Statistics early in each month. The September 2026 release showed the index at 120.43 (2020=100), up 0.3% on the month and 2.9% on the year. Inflation excluding food and energy was 2.8%. The chart below uses the trend table in that release.
Korean CPI, percent change year on year (February to September 2026)
Headline and excluding food and energy, not seasonally adjusted, against the Bank of Korea 2% inflation target.
Takeaway: headline inflation has been at or above 2.8% since May and core jumped to 3.4% in August, which matters for USD/KRW through Bank of Korea rate expectations. Both KOSIS and FXMacroData carry these numbers; the difference is when and how your system learns them.
Release Timing: 08:00 KST And What The API Records
The September 2026 CPI release was marked "Embargo 8:00 A.M., October 2, 2026", with a 7:30 A.M. release time on the press release header. That makes 08:00 KST (23:00 UTC the previous evening) the moment the market can act on the number, and it falls during the Asia morning, when USD/KRW onshore trading opens at 09:00 KST. The FX sessions view shows how that overlaps with the other markets.
KOSIS is not built to tell you that time. Its data rows carry the reference period and a date-level last-changed field. That is fine for research tables but does not let a backtest tell whether a number was known at 08:00:01 or late that afternoon. We have not measured how soon each KOSIS table updates after an 08:00 release and make no claim about it. The point is narrower: the API does not record a publication timestamp you can join to tick data.
FXMacroData reads the same official release. Its KRW CPI series is built from the official Statistics Korea index, as republished in the Bank of Korea's statistics database. It adds the second-level announcement_datetime, the prior value, and source metadata to each row. The target is one second from publication, and the measured record by source is public on the release-speed page, including every release that took longer.
Korean CPI release morning (KST)
07:30 press copy
Release date printed on the press release; the figure stays under embargo.
08:00 embargo lifts
The number becomes public. This is the event time an FX model needs.
FXMacroData row
Value, prior, source, and announcement_datetime to the second, with measured pickup published.
KOSIS table
The table carries the new month with a date-level last-changed stamp, plus the full sub-item detail.
09:00 onshore open
USD/KRW onshore trading opens with the print already known.
Takeaway: for KRW event studies, you need to record the 08:00 publication moment. FXMacroData stores it on every row; KOSIS stores the statistics.
Rate Limits, Formats, And Licence Terms
The KOSIS service-introduction page lists the same limits for each service: up to 200 calls a minute and up to 40,000 cells per request. Bulk data has a higher ceiling of 200,000 cells in XLS format. Statistical data comes as SDMX or JSON, survey and table descriptions as XML or JSON, and integrated search as JSON. That is generous for a research pipeline that refreshes a few dozen tables, and tight only if you try to mirror large parts of the database.
The terms of use are permissive and clear. Commercial use of the statistics is allowed, with exceptions announced on the site. Users must show the source and follow the statistics usage guidelines. Passing the received data unchanged to third parties for a fee without your own service platform is prohibited. The terms renew yearly. They also place responsibility on the user for checking notices about changes, and they let the provider pause or end the service for maintenance or replacement. For a production trading system, that last point means building your own monitoring.
FXMacroData's terms are commercial in the other direction: every plan covers your own and your team's internal use, and showing the data to outside users is handled by the self-serve Commercial Redistribution add-on at $10/month per block of up to 100 measured users, or up to 1,000 public-feed followers.
Why Korean Data Moves USD/KRW
The won is one of the most trade-sensitive major Asian currencies. Three Korean data streams feed a typical USD/KRW process. First, CPI drives Bank of Korea rate expectations against the Federal Reserve. Second, monthly trade and the balance of payments show dollar supply from exporters. Third, the Korea Customs Service publishes early export tallies for the first 10 and first 20 days of each month. Global investors read those partial-month export figures, which are heavy in semiconductors, as a bellwether for world demand and the chip cycle. The tally comes from customs, not KOSIS, so a pure KOSIS build does not capture it on release day.
The useful view joins those Korean releases to US data, bond yields, and the pair itself. That is the work FXMacroData is built for. Korean GDP, CPI, trade balance, current account, policy rate, and bond yields share one schema with US CPI and payrolls, so a single loop can rank surprises across the pair's two sides. The South Korea data hub lists every KRW series and its most recent release.
Use-case storyboard: a USD/KRW release week
Plan
Pull the KRW and USD release calendar; flag the 08:00 KST CPI and the next US CPI.
Capture
Store the KRW CPI row with its second-level timestamp and prior value as it lands.
Deepen
If the surprise is large, open KOSIS for the sub-item table to see whether energy, food, or services drove it.
Act
Feed the row to a model, alert, or MCP-connected assistant that compares it with US data and the USD/KRW move.
Takeaway: FXMacroData handles the timed, cross-currency part of the week; KOSIS is the place to go when you need to explain the number in detail.
API Example
FXMacroData sends the key in a header and uses the same path pattern for every currency. This request returns the Korean CPI announcement history:
curl -H "X-API-Key: YOUR_API_KEY" "https://api.fxmacrodata.com/v1/announcements/krw/inflation"Example response (abridged, illustrative values for the September 2026 print):
{
"currency": "KRW",
"indicator": "inflation",
"source": "Statistics Korea",
"data": [
{
"date": "2026-09-30",
"val": 2.9,
"val_mom": 0.3,
"previous_value": 3.1,
"previous_date": "2026-08-31",
"announcement_datetime": 1790895600,
"announcement_datetime_local": "2026-10-02T08:00:00+09:00"
}
]
}FXMacroData row
announcement_datetime: the 08:00 KST publication, in epoch seconds.valandprevious_value: ready for surprise maths.- Same fields for USD, JPY, EUR, and every other currency.
KOSIS row
PRD_DE: the reference period, such as 202609.DT: the value, returned as a string.LST_CHN_DE: the last-changed date, with no time of day.
Takeaway: both rows hold the same official number. Only one of them can anchor an intraday event study without extra work.
Where KOSIS Is Genuinely Stronger
For Korean statistics on their own terms, KOSIS is the better tool. It is the official portal, it is free, and the terms allow commercial use with attribution. It goes far deeper than any FX product: CPI by expenditure category and individual items, regional series, survey methodology, and table metadata that explains exactly what each number measures. Researchers who need a long, authoritative history or a precise citation should start there. SDMX output and English labels in the responses also make it workable for international data teams, despite the Korean-language portal.
KOSIS is also the right place to explain a print. When September 2026 transport prices rose 7.7% year on year, that detail was in the official release tables. A headline FX feed tells you the number moved; the KOSIS tables tell you why.
Where FXMacroData Is Stronger
FXMacroData is stronger once Korea is one leg of a currency decision. You do not maintain table and item codes or watch a notice board for rebasing. Every release row carries a second-level publication time with a public speed record. Korean data shares a schema with 21 other currencies, and the same key unlocks the calendar, FX spot rates, COT positioning, bond yields, central-bank press releases, and an MCP server that lets an AI trading assistant query it directly.
FXMacroData pros
- Second-level release timestamps with a public speed record.
- One schema across 22 currencies, including KRW and USD.
- Calendar, pair dashboards, exports, and MCP in one plan.
FXMacroData tradeoffs
- Paid plan required for KRW; the free tier is USD only.
- Headline series, not every sub-item or region.
- Not the official publisher; cite the agency for official work.
KOSIS pros
- Free, official, commercial use permitted with attribution.
- Very deep table coverage and survey metadata.
- SDMX and JSON output; English labels in responses.
KOSIS tradeoffs
- Korean-language portal, guide, and application flow.
- No publication timestamp on data rows.
- Korea only; table codes to maintain; no FX or calendar layer.
Takeaway: KOSIS is the stronger statistics source. FXMacroData is the stronger choice when KRW data has to drive a timed, multi-currency trading process.
Decision Matrix
| Reader | Better fit | Why |
|---|---|---|
| Korea economist or academic | KOSIS | Official tables, sub-items, regional detail, and metadata at no cost. |
| USD/KRW discretionary trader | FXMacroData | Calendar, pair dashboard, and timed KRW and USD releases in one place. |
| Quant building KRW event studies | FXMacroData | Second-level publication timestamps make the 08:00 KST event joinable to tick data. |
| Asia multi-currency desk | FXMacroData | KRW, JPY, CNY, TWD, THB, MYR, and USD in one schema instead of six national APIs. |
| AI or automation builder | FXMacroData | MCP server, OpenAPI spec, and English docs reduce adapter work. |
| Research team explaining a CPI surprise | Both | FXMacroData for the timed headline; KOSIS for the item-level breakdown. |
Recommendation
Use KOSIS if your work is Korean statistics: you need item-level CPI, regional data, survey metadata, or an official citation, and you can live with a Korean-language portal and date-level update stamps. It is free, authoritative, and permits commercial use with attribution.
Use FXMacroData if Korean data is an input to a currency decision. The Individual plan at $50/month (14-day trial) covers KRW alongside 21 other currencies, with second-level release timestamps, a calendar, pair dashboards, and MCP access. Compare plans and pricing, review the API reference, and check the measured pickup times on the release-speed record before you commit.
Bottom line: KOSIS is the official, free, and deepest source of Korean statistics. FXMacroData reads the same official releases and turns them into timestamped, cross-currency rows that a USD/KRW model, dashboard, or AI assistant can act on. Choose by the job: statistics project, KOSIS; trading workflow, FXMacroData.
Sources Checked
KOSIS OpenAPI service introduction
Services, formats, 200 calls/minute, 40,000 cells per request.
View sourceKOSIS OpenAPI home and notices
Table counts and 2026 notices on HTTPS and per-minute limits.
View sourceMinistry of Data and Statistics CPI release
September 2026 CPI, 08:00 embargo, trend table used in the chart.
View sourceTakeaway: every KOSIS limit and term quoted here comes from the official portal pages above, checked in October 2026.
Related Reading
- Bank of Korea policy and the KRW: the rate path that Korean CPI feeds into.
- Semiconductor sales and the Korean won: why chip exports matter for the currency.
- FXMacroData vs ECB Data Portal API: the same official-source comparison for the euro area.
- Best macroeconomic data APIs in 2026: a wider view of the options.
