California Tax Appeals README

@pipeworx/california-tax-appeals

California Office of Tax Appeals (OTA) opinions — search by statute citation or topic across 4,254 issued opinions (precedential and nonprecedential, 2018–2026), or retrieve one specific opinion's full text and cited provisions by its OTA citation.

Part of Pipeworx — an MCP gateway connecting AI agents to 1784+ live data sources. This is an independent, unofficial integration — not affiliated with, endorsed by, or published by the upstream provider.

Tools

• ca_ota_search(query, category?, precedential_only?, limit?) — find opinions by topic keyword (e.g. "residency", "innocent spouse") or statute citation (e.g. "17014"). Returns citation, title, date, tax category (Franchise/Income Tax vs Business Tax), precedential status, and OTA's own one-line "Issues" tag, most recent first. Searches listing metadata only (title + Issues tag) — OTA publishes no full-text search.
• ca_ota_opinion_lookup(cite) — fetch one opinion by its OTA citation (e.g. "2020-OTA-263P"), live. Returns the full opinion text, exact date issued, OTA's internal case number, tax category, precedential status, and statute citations found in the text (best-effort R&TC/Cal. Code Regs./IRC).

Auth

Keyless. No signup — this is a read of OTA's own published, public opinions.

Data sources

• https://ota.ca.gov/opinions/ — the landing page this pack's bake script (scripts/bake-index.mjs) scrapes to discover every per-year and all-precedential opinion listing page.
• Per-year listing pages, e.g. https://ota.ca.gov/2024-franchise-income-tax-opinions/ and https://ota.ca.gov/2024-business-tax-opinions/ — every opinion issued that year/category, precedential and nonprecedential alike.
• All-years precedential pages, e.g. https://ota.ca.gov/all-precedential-franchise-and-income-tax-opinions/ — fetched too, to cover any year a per-year page might omit.
• Individual opinion PDFs (e.g. https://ota.ca.gov/wp-content/uploads/sites/54/2021/06/19064883_Mazer_Opinion_Revised_090420_P.pdf) — fetched live per ca_ota_opinion_lookup call, text extracted with shared/src/pdf-text.ts.

Shape and known limits

ota.ca.gov (WordPress/WP Engine, TablePress tables) has no JSON search backend — every opinion-listing page renders its whole table as static HTML; the DataTables plugin paginates it client-side only (confirmed live: every DataTable('#tablepress-NNN', {...}) init call on these pages carries no ajax/serverSide option). So this pack bakes a citation-indexed table from OTA's own listing pages rather than proxying a search endpoint — see scripts/bake-index.mjs's own header for exactly which pages it fetches and how it discovers new years without a code change.

Per root CLAUDE.md's standing rule on baked data (fleet #2754), the index (4,254 rows, ~1.6MB) lives in KV, never gateway module scope — uploaded by scripts/sync-gateway-static-json.mjs and injected as args._bakedIndex (registered in workers/gateway/src/pack-baked-indexes.json).

ca_ota_search searches listing metadata only — the opinion title and OTA's own one-line "Issues" tag (e.g. "Residency (R&TC 17014; Reg. 17014)"). A statute cited only in an opinion's full body text, and not in its Issues tag, will not be found by search; use ca_ota_opinion_lookup on a specific citation to read that opinion's full text.

"Case number" = OTA's own citation (e.g. "2020-OTA-263P", the trailing "P" marking Precedential status) — what OTA itself cites opinions by, not the PDF's internal 8-digit "OTA Case No." docket number (which only appears inside the PDF body and is surfaced in ca_ota_opinion_lookup's response as ota_case_no, read live, not indexed). The trailing precedential marker may be omitted when looking up a citation; ca_ota_opinion_lookup tries both forms before reporting a miss. Some opinions carry other OTA suffix letters (e.g. "...SCP" for a Summary Conclusion of Proceedings) — those must be typed exactly.

A transport failure (non-200) or a listing/PDF mismatch (a 404 for a citation the index says has a PDF) throws loudly naming the upstream status — never a silent found: false. A PDF that exists but fails to extract (rare; e.g. a scanned, non-text PDF) degrades to full_text: null with a warning rather than throwing, since the opinion's metadata (title, date, decision type, Issues tag) is still valid from the listing page.

Re-run node scripts/bake-index.mjs > src/ca-ota-index-data.ts periodically to pick up newly issued opinions — it refuses to write a suspiciously small index (fewer than 2,000 opinions or fewer than 15 listing pages fetched), so a silent partial bake from an OTA markup change is caught rather than shipped.

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "california-tax-appeals": {
      "url": "https://gateway.pipeworx.io/california-tax-appeals/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/california-tax-appeals/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Both URLs reach the same gateway and the same 1784+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

No MCP client? Call it over HTTP

curl -X POST https://gateway.pipeworx.io/v1/tools/ca_ota_search \
  -H 'Content-Type: application/json' \
  -d '{"query":"residency"}'

No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/ca_ota_search. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.

Standalone (no gateway account)

This package also runs as a local stdio MCP server — no Pipeworx account, no gateway round-trip:

{
  "mcpServers": {
    "california-tax-appeals": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-california-tax-appeals"]
    }
  }
}

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-california-tax-appeals

It speaks MCP over stdin/stdout and answers initialize/tools/list/tools/call for only this pack's tools — none of the shared meta-tools the gateway connection above adds. Same source, same tools, no ask_pipeworx routing.

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:

ask_pipeworx({ question: "your question about California Tax Appeals data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides
• pipeworx.io

License

MIT