Nomina
Official# Nomina
Markets research for agents.
Four read-only tools, no API key, no account, no trading access:
- **`search_markets`** — find ticker symbols and dated headlines for a company, asset, sector,
or market topic.
- **`research_asset`** — quote, period return, drawdown and volatility, dated price history,
and related news for one symbol, over a named period or an exact date window.
- **`compare_assets`** — align 2–6 symbols on shared observation dates, in local currencies,
with an explicit adjusted-vs-raw basis.
- **`market_overview`** — period returns for major indexes, rates, the dollar, gold, oil,
BTC, ETH and EURUSD.
## Install
**Claude Desktop:** Settings → Extensions → Advanced settings → Extension Developer →
Install Extension… → select `Nomina.mcpb` from a
[release](https://github.com/nomina-xyz/nomina-mcp/releases) (or build it yourself, see below). Requires a Claude Desktop
release with MCPB v0.4 UV-runtime support and internet access on first install.
**Any other MCP client (Claude Code, etc.):** point it at this as a local stdio server:
```json
{
"mcpServers": {
"nomina": {
"command": "uv",
"args": ["run", "--frozen", "--no-dev", "--directory", "/path/to/nomp", "server.py"]
}
}
}
```
**Self-hosted (Streamable HTTP):** `docker run --rm -p 8000:8000 ghcr.io/nomina-xyz/nomina-mcp:1.3.0`,
then connect to `http://localhost:8000/mcp`. Details in the docs below.
## Docs
Install guides, tool reference, privacy policy, and data-source limits:
https://nomina-xyz.github.io/nomina-mcp/ — see especially
[Data sources and limits](https://nomina-xyz.github.io/nomina-mcp/data-sources/).
## Build from source
Requires [uv](https://docs.astral.sh/uv/).
```sh
uv sync
uv run python scripts/build_bundle.py # writes dist/Nomina.mcpb
uv run pytest -q # 14 tests
uv run ruff check .
```
## License
Code in this repository is MIT-licensed (see `LICENSE`).
TDQS
Scored across 4 tools
Each tool has a clearly distinct role: search_markets discovers symbols, research_asset covers a single asset, compare_assets handles multi-asset comparisons, and market_overview provides a macro snapshot. There is no meaningful overlap between the tools.
Three tools follow a consistent verb_noun pattern (search_markets, research_asset, compare_assets), while market_overview deviates by being noun-only. The naming is still clean, lowercase, and readable, so the inconsistency is minor.
Four tools is an appropriate, well-scoped size for a market data research server. Each tool earns its place and covers a distinct part of the research workflow without unnecessary redundancy.
The tool set covers the core research lifecycle: discover symbols, research one asset, compare multiple assets, and get a broad market snapshot. Limitations like no financial statements or full articles are explicitly scoped out rather than missing functionality.