# PlatPhorm Data Studio

PlatPhorm Data Studio is the structured-data transformation workspace at https://data.platphormnews.com.

## Product purpose

Drop a file, immediately understand it, transform or query it locally, preview the result, and export it. The site remains a data tool; platform observability and discovery support that purpose without replacing it.

Publisher: PlatPhormNews (https://platphormnews.com). Product and discovery content last reviewed 2026-07-23.

## Local processing boundary

- DuckDB-Wasm runs inside a dedicated Web Worker.
- Native structured files use browser file handles for streaming/random-access reads where possible.
- The database prefers OPFS-backed persistence and reports an honest in-memory fallback.
- Source files and source rows are not uploaded to the application server.
- Local imports and downloads have no application byte cap. CSV, TSV, JSON, NDJSON, and Parquet stream from browser file handles; parser-based formats remain bounded by available browser memory.
- A generated export is transferred only after the user explicitly chooses a compatible PlatPhorm destination, supplies PLATPHORM_API_KEY, and selects Publish.
- The protected publish route forwards the selected export to a fixed allowlisted service, requires durable destination evidence, and does not store the export in Data Studio.

## Import and detection

Supported imports: CSV, TSV, JSON, NDJSON, XML, YAML, XLSX, Parquet, Arrow IPC, SQLite, Apple/iOS vCard, and Google Contacts CSV. Contact imports normalize names, email addresses, phone numbers, organizations, addresses, websites, notes, groups, birthdays, and identifiers. Google source rows remain available in the normalized `source_record` field; vCard photos and embedded binary attachments are intentionally ignored.

On import the studio derives row count, schema, column types, encoding where applicable, delimiter, missing values, duplicate rows, malformed sampled rows, ranges, top values, numeric histograms, date ranges, file size, workspace memory, potential issues, and deterministic suggested fixes.

## Transform and query

Point-and-click operations: filter, sort, rename, keep/drop columns, type conversion, deduplication, null cleanup, flatten objects/lists, pivot, calculated columns, join, union, split, schema comparison, and row-level dataset diff.

The SQL console runs read-only query previews locally in DuckDB-Wasm and keeps query history in browser-local storage.

## Export

Supported exports: CSV, TSV, JSON array, NDJSON, Parquet, Arrow IPC, SQLite, Markdown, HTML, SQL CREATE/INSERT statements, iOS-compatible vCard, and Google Contacts CSV. The active visual transformation pipeline is included in exports.

Optional protected publishing is format-aware:

- PlatPhorm Sheets: CSV, TSV, JSON, Markdown, HTML, and Google Contacts CSV
- PlatPhorm Docs: CSV, TSV, JSON, NDJSON, Markdown, HTML, SQL, vCard, and Google Contacts CSV
- PlatPhorm OpenContent: CSV, TSV, JSON, NDJSON, Markdown, HTML, SQL, vCard, and Google Contacts CSV
- PlatPhorm Files: every supported export format, including Parquet, Arrow, and SQLite

Sheets, Docs, and OpenContent publishing uses a 4 MiB JSON/document relay. PlatPhorm Files switches larger generated exports to a protected short-lived direct upload up to 2 GiB and indexes the object only after byte-count and S3 SHA-256 verification. Local downloads have no application size cap.

## API and MCP

- Public API documentation: https://data.platphormnews.com/api/docs
- OpenAPI 3.1: https://data.platphormnews.com/openapi.yaml
- Public declared compliance: GET https://data.platphormnews.com/api/v1/compliance
- Protected live compliance check: POST https://data.platphormnews.com/api/v1/compliance
- Protected opt-in publishing: POST https://data.platphormnews.com/api/v1/publish
- MCP metadata: GET https://data.platphormnews.com/api/mcp
- MCP JSON-RPC: POST https://data.platphormnews.com/api/mcp

The MCP surface is intentionally read-only and real: it exposes supported formats, implemented transforms, compatible publish destinations, the local workflow, declared route compliance, documentation resources, and cleanup/comparison prompts. It does not accept source data or claim server-side execution of browser-local datasets.

Protected actions accept either `Authorization: Bearer $PLATPHORM_API_KEY` or `X-PlatPhorm-API-Key: $PLATPHORM_API_KEY`.

## Trust

Publisher, corrections, contact, AI-use, and source-attribution policies are public and crawlable:

- https://data.platphormnews.com/policies/corrections
- https://data.platphormnews.com/policies/contact
- https://data.platphormnews.com/policies/ai-use
- https://data.platphormnews.com/policies/source-attribution

Web dashboard, public-safe discovery, browser-based operations, trusted-domain discovery, standard route compliance, Vercel metadata capture, trace inspection, and agentic workflow discovery are intentionally supported for public read-only debugging and operator workflows. Mutating, administrative, ingestion, replay, fork, remediation, deployment, sync, test-triggering, reporting, and write actions require PLATPHORM_API_KEY.