# styx-search

**Search 500,000+ open-access papers, standards, textbooks, and legal documents — right from your terminal or your AI agent.**

`styx-search` is a zero-configuration CLI and [Model Context Protocol (MCP)](https://modelcontextprotocol.io) server for the [Styx Open Knowledge Archive](https://styx-iota.vercel.app). Every source in the archive is open access or public domain — no paywalls, no accounts required to search.

- **For humans** — search, export citations in 11 formats (BibTeX, APA, MLA, Chicago, IEEE, Harvard, RIS, Bluebook, OSCOLA, ISO 690, ANSI Z39), and download PDFs to disk with one command.
- **For AI agents** — a stdio MCP server with five tools, so Claude, Cursor, Windsurf, OpenCode, or any MCP client can search, resolve identifiers (RFC, DOI, arXiv, PMID, PMCID, NIST/FIPS, W3C, US case citation), and save PDFs into your workspace.
- **100% self-contained** — talks only to the public HTTPS API. No API keys, no sign-up, no internal infrastructure.

---

## Install

```bash
pip install styx-search
```

Or from source:

```bash
git clone https://github.com/0x923041-dotcom/styx-cli.git
cd styx-cli
pip install .
```

Requires Python 3.8+.

---

## Quick start

```bash
# Search — supports source:/lang:/category: directives
styx search "attention is all you need" --limit 5
styx search "source:nist zero trust" --category "Computer Science & Tech"
styx search "source:eurlex AI act" --json
styx search "meditations marcus aurelius" --markdown   # for Obsidian/Notion
styx search "diffusion models" --offset 20            # paginate past the first page

# Resolve a known identifier straight to its document(s)
styx resolve "RFC 2119"
styx resolve "doi:10.1145/3290605.3300333"
styx resolve "2604.08499"                              # arXiv
styx resolve "PMID 41961061"                           # PubMed
styx resolve "PMC10601397"                             # PubMed Central
styx resolve "NIST FIPS 199"
styx resolve "410 U.S. 113"                            # US case citation (Supreme Court)
styx resolve "20-5364"                                 # federal docket
styx resolve "RFC 2119" --json

# Download the original PDF
styx download 297465 -o paper.pdf          # by document ID
styx download "https://arxiv.org/pdf/2604.21816v1.pdf"   # by URL

# Citations — 11 formats: bibtex, apa, mla, chicago, ieee, harvard,
# ris, bluebook, oscola, iso690, ansiz39
styx cite 7288 --format bibtex
styx cite 7288 --format apa
styx cite 7288 --format iso690

# Document metadata
styx info 2844

# License key (Pro) + quota
styx auth STYX-XXXX-XXXX-XXXX
styx quota

# Version
styx --version

# AI agent server
styx mcp
```

### Example

```text
$ styx search "source:nist zero trust" --limit 3

Styx - 3 results for "source:nist zero trust"
  #  Year  Title                                       Author                  Source           PDF
  1  n.d.  NIST SP 1800-35: Implementing a Zero Trust  Scott Rose (NIST); …    NIST             [x]
  2  n.d.  NIST SP 800-207: Zero Trust Architecture     Scott Rose (NIST); …    NIST             [x]
  3  n.d.  NIST SP 800-207A: A Zero Trust Architecture Ramaswamy Chandramo…   NIST             [x]

  Tip: styx download 2844 saves the PDF, styx cite 2844 exports a citation.
```

Query words are highlighted in gold in the title column. The `PDF` column shows
`[x]` when a direct PDF link is available; `[ ]` rows still surface the source
URL via `styx info <id>`.

### Search directives

The CLI understands the same directives as the Styx web app — they are stripped
from the query and applied as real filters:

| Directive | Example | Effect |
|-----------|---------|--------|
| `source:` / `src:` | `source:nist` | Restrict to one or more sources (`source:nist source:owasp`) |
| `lang:` / `language:` | `lang:it` | Restrict to a language (ISO code) |
| `category:` / `cat:` | `cat:"Law & Regulation"` | Restrict to a category tab |

Directives can be combined freely: `styx search "cat:\"Computer Science & Tech\" source:nist lattice cryptography"`.

### Resolve

`styx resolve` recognises several identifier shapes and returns a typed
response (`kind: rfc | doi | arxiv | pmid | pmcid | nist | w3c | case` plus
fallback `null` for plain text). Unrecognised identifiers degrade to a regular
search so the command never fails silently.

---

## Model Context Protocol (MCP)

The `styx mcp` command starts a stdio JSON-RPC MCP server. It is dependency-free
(no MCP SDK required) and works with any MCP client.

### Tools

| Tool | Description |
|------|-------------|
| `styx_search` | Search the archive (`query`, optional `limit` 1–20, optional `category`). Returns structured results with titles, authors, years, source, snippet, and `direct_pdf_url` when available. |
| `styx_resolve` | Resolve a known identifier — RFC (`RFC 2119`), DOI (`doi:10.1145/3290605.3300333`), arXiv (`2604.08499`), PubMed ID (`PMID 41961061`), PubMed Central ID (`PMC10601397`), NIST/FIPS designation (`NIST FIPS 199`), W3C spec shortname (`w3c_digital-credentials`), or US legal citation / docket (`410 U.S. 113`, `20-5364`) — straight to its document(s). Unrecognised identifiers fall back to a text search. |
| `styx_download_pdf` | Download the original open-access PDF for a Styx document ID or a direct PDF URL to a local path. |
| `styx_get_citation` | Get a citation for a document ID in any of 11 formats (BibTeX, APA, MLA, Chicago, IEEE, Harvard, RIS, Bluebook, OSCOLA, ISO 690, ANSI Z39). |
| `styx_get_document` | Full metadata (title, author, source, category, license, year, language, URL, direct PDF URL, description) for a document ID. |

### Claude Desktop

Add to `claude_desktop_config.json`:

- **macOS:** `~/Library/Application Support/Claude/claude_desktop_config.json`
- **Windows:** `%APPDATA%\Claude\claude_desktop_config.json`

```json
{
  "mcpServers": {
    "styx": {
      "command": "styx",
      "args": ["mcp"]
    }
  }
}
```

### Cursor

`Settings → Cursor Settings → MCP → Add new MCP server`:

```json
{
  "mcpServers": {
    "styx": {
      "command": "styx",
      "args": ["mcp"]
    }
  }
}
```

### OpenCode / Windsurf / Antigravity

The same JSON shape works in any MCP-aware client — point it at `styx mcp`
and the five tools above are exposed automatically.

After adding, restart the client and you can ask, for example:

> *"Find the NIST post-quantum encryption standard and download the PDF into my workspace."*

### Manual smoke test

```bash
echo '{"jsonrpc":"2.0","id":1,"method":"tools/list"}' | styx mcp
```

---

## Fair use

Every source in Styx is open access or public domain, and searching is free.
To keep the archive open for everyone, a trial allowance applies per client
(and then a rate limit). When a limit is reached, the CLI returns a clear,
non-blocking message:

```text
Free trial allowance exhausted. Add a license key to continue at higher
limits: docs at https://styx-iota.vercel.app or `styx auth <key>`.
```

Add a license key once with `styx auth <key>` — it is stored in
`~/.styx/config.json` and sent with every request. No account or sign-up is
required to search. `styx quota` reports your current tier (`public` / `pro` /
`institutional`).

---

## Configuration

State lives in `~/.styx/config.json`:

```json
{
  "api_base": "https://styx-iota.vercel.app",
  "license_key": "",
  "default_limit": 10,
  "default_citation_format": "bibtex"
}
```

Override the API base (for staging, self-hosted mirrors, or a local proxy) by
editing `api_base`. Set `default_limit` to change the search page size; set
`default_citation_format` to any of the 11 supported format names.

---

## Development

```bash
pip install -e ".[test]"     # or: pip install -e . && pip install pytest
pytest tests/                # 41 offline tests (mocked HTTP, no network)
```

The test suite is fully offline — `httpx` is patched with a `MockTransport`
and `StyxClient` is stubbed where it composes other services. No test makes a
real network call.

---

## License

MIT — see [LICENSE](LICENSE).

Built on the [Styx Open Knowledge Archive](https://styx-iota.vercel.app) — 500+
living sources: arXiv, NIST, OWASP, CISA, PubMed/PMC, MIT OpenCourseWare,
EUR-Lex, Project Gutenberg, and more.