datacore-cli

Vietnamese financial and alternative data, from your terminal.

Website · Docs · API Reference · Changelog

---

datacore is the official CLI for DataCore — Vietnam's financial and alternative data platform. Search the catalog, preview datasets, and download production-grade data to CSV or Parquet, all without leaving your terminal.

pip install datacore-cli
export DATACORE_API_KEY=dc_...

datacore search "VN30 fundamentals"
datacore get equity.vn30.daily --start 2024-01-01 --output vn30.parquet

---

Installation

pip install datacore-cli

Requires Python 3.10+. The CLI is a thin wrapper around the DataCore Python SDK; it is installed automatically.

Shell completions (optional)

# Bash
datacore --install-completion bash

# Zsh
datacore --install-completion zsh

# Fish
datacore --install-completion fish

---

Authentication

Get your API key at datacore.vn/dashboard.

# Option 1 — environment variable (recommended for CI/scripts)
export DATACORE_API_KEY=dc_...

# Option 2 — persistent config file
datacore config set api_key dc_...

The config file lives at ~/.config/datacore/config.json on Linux/macOS and %APPDATA%\datacore\config.json on Windows.

---

Commands

Discovery

Command	Description
datacore search QUERY	Hybrid keyword + semantic search across the entire catalog
datacore domains	List all data domains (equity, macro, alternative, ...)
datacore products DOMAIN	List products within a domain
datacore datasets	List datasets — filter with --product PRODUCT_ID

Inspection

Command	Description
datacore meta DATASET_ID	Full dataset metadata as JSON
datacore schema DATASET_ID	Column-level schema: name, type, description
datacore sample DATASET_ID	Stream preview rows (default: 10, change with -n N)

Download

Command	Description
datacore get DATASET_ID	Download to CSV or Parquet — required: --output FILE

Options for get:

--start DATE    Start date (ISO 8601, e.g. 2024-01-01)
--end   DATE    End date   (ISO 8601)
--output FILE   Destination file (.csv or .parquet)
--format FMT    Override output format

Configuration

Command	Description
datacore config show	Dump the active config (file + env merged)
datacore config set KEY VALUE	Persist a setting (api_key, base_url, output_dir, default_format)
datacore config get KEY	Read a single setting

Integrations

Command	Description
datacore mcp	Write the DataCore entry into Claude Desktop's MCP config
datacore version	Print the installed CLI version

---

Output formats

Pass --format to any tabular command:

Format	Flag	Notes
Table (default)	--format table	Rich-formatted, colour terminal output
JSON	--format json	One JSON object per line (NDJSON)
CSV	--format csv	Comma-separated, UTF-8
Parquet	--format parquet	Columnar binary — best for large datasets

---

Examples

Search and download VN30 OHLCV

# Discover available equity datasets
datacore search "VN30 OHLCV"
datacore schema equity.vn30.daily

# Download two years of daily data
datacore get equity.vn30.daily \
  --start 2022-01-01 \
  --end   2023-12-31 \
  --output vn30_2y.parquet

Explore macroeconomic indicators

datacore domains
datacore products macro
datacore datasets --product macro-vn
datacore sample macro.vn.cpi -n 20

Pipe JSON into jq

datacore meta equity.vn30.daily --format json | jq '.fields[].name'

Use in a Python script

import subprocess, json

result = subprocess.run(
    ["datacore", "meta", "equity.vn30.daily", "--format", "json"],
    capture_output=True, text=True
)
meta = json.loads(result.stdout)

---

Python SDK

datacore-cli is built on the DataCore Python SDK, which you can use directly for programmatic access:

from datacore import Datacore

client = Datacore(api_key="dc_...")
df = client.get_data("equity.vn30.daily", columns=["date", "close_price"])["data"]

See the SDK docs for the full API reference.

---

Claude Desktop / MCP integration

datacore-cli ships a built-in MCP server that connects Claude Desktop to your DataCore account. Run:

export DATACORE_API_KEY=dc_...
datacore mcp

This writes the DataCore entry into claude_desktop_config.json. Restart Claude Desktop — you can then ask Claude to query DataCore datasets directly in conversation.

---

Contributing

Contributions are welcome! Please read CONTRIBUTING.md before opening a pull request.

Quick start:

git clone https://github.com/DataCore-VietNam/datacore-cli
cd datacore-cli
pip install -e ".[dev]"
pre-commit install
pytest

---

License

MIT © DataCore Vietnam