mcp-public-intel
# mcp-public-intel
[](LICENSE)
[](https://www.python.org/downloads/)
[](#available-tools)
[](#quick-start)
**Zero-API-key Model Context Protocol (MCP) server for live public intelligence.**
Plug real-time SEC EDGAR financials & Form D startup raises, Greenhouse/Lever/Ashby engineering job boards, Shopify store catalogs, Substack public posts, and Podcasting 2.0 RSS transcripts directly into **Claude Desktop, Cursor, Antigravity, Windsurf, or Codex** in 10 seconds — with **zero API keys**, **one dependency (`httpx`)**, and **<100ms stdio startup**.
---
## Quick Start (Zero API Keys)
### Claude Desktop / Cursor / Antigravity / Windsurf (`mcpServers`)
Paste into your MCP config (`claude_desktop_config.json`, `.cursor/mcp.json`, or Antigravity MCP config):
```json
{
"mcpServers": {
"public-intel": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/pecopeko/mcp-public-intel",
"mcp-public-intel"
]
}
}
}
```
### Or Install Locally with `pip`
```bash
pip install "git+https://github.com/pecopeko/mcp-public-intel"
mcp-public-intel --list-tools
```
Then configure your client with `"command": "mcp-public-intel"` (or `"command": "python3", "args": ["-m", "mcp_public_intel"]`).
---
## Available Tools (6 Live Zero-Key Sources)
| Tool | Required Inputs | Optional Inputs | Official Public Data Source |
| :--- | :--- | :--- | :--- |
| **`sec_company_financials`** | `ticker_or_cik` | `contact_email` | `data.sec.gov` submissions + ASC 606 XBRL `companyfacts` (separates annual 10-K vs quarterly 10-Q revenue, net income, operating income, cash flow, assets, cash, shares) |
| **`sec_form_d_funding_search`** | *(none)* | `query`, `date_from`, `date_to`, `limit` | `efts.sec.gov/LATEST/search-index` (recent Form D & D/A private startup funding filings) |
| **`ats_company_jobs`** | `provider`, `company_slug` | `remote_only`, `keyword`, `limit` | Public Greenhouse, Lever, and Ashby JSON APIs (normalized remote/hybrid, salary, seniority, tech tags) |
| **`shopify_store_catalog`** | `store_url` | `limit`, `collection_handle` | Public Shopify `/products.json` & `/collections.json` (variants, min/max prices, discount %, stock status) |
| **`substack_public_posts`** | `publication` | `limit`, `only_free` | Public Substack `/api/v1/archive` + `/feed` RSS to clean Markdown (honestly flags paywalled previews; zero bypass) |
| **`podcast_rss_transcripts`** | `query_or_rss_url` | `limit` | Apple iTunes Lookup/Search API + Podcasting 2.0 RSS (`<itunes:owner>` host emails + VTT/SRT/JSON transcripts) |
---
## Sample JSON Outputs
### 1. `sec_company_financials` (`{"ticker_or_cik": "AAPL"}`)
```json
{
"source": "sec_edgar_xbrl",
"status": "ok",
"cik": "0000320193",
"entity_name": "Apple Inc.",
"tickers": ["AAPL"],
"xbrl_status": "extracted",
"annual_financials": {
"revenue": {
"tag": "RevenueFromContractWithCustomerExcludingAssessedTax",
"val": 391035000000,
"unit": "USD",
"start": "2023-10-01",
"end": "2024-09-28",
"form": "10-K",
"fp": "FY"
},
"net_income": { "tag": "NetIncomeLoss", "val": 93736000000, "unit": "USD", "form": "10-K" },
"shares_outstanding": { "tag": "EntityCommonStockSharesOutstanding", "val": 15115823000, "unit": "shares", "form": "10-K" }
},
"quarterly_financials": {
"revenue": { "tag": "RevenueFromContractWithCustomerExcludingAssessedTax", "val": 94930000000, "unit": "USD", "form": "10-Q", "period_days": 90 }
}
}
```
### 2. `sec_form_d_funding_search` (`{"query": "robotics", "limit": 5}`)
```json
{
"source": "sec_efts_form_d",
"status": "ok",
"total_hits": 412,
"returned_count": 5,
"filings": [
{
"entity_name": "Acme Robotics Inc.",
"form": "D",
"file_date": "2026-11-18",
"primary_cik": "0002012345",
"biz_locations": ["San Francisco, CA"],
"edgar_url": "https://www.sec.gov/Archives/edgar/data/2012345/000201234526000001/primary_doc.xml"
}
]
}
```
### 3. `ats_company_jobs` (`{"provider": "ashby", "company_slug": "linear", "remote_only": true, "limit": 5}`)
```json
{
"source": "ats_ashby",
"status": "ok",
"provider": "ashby",
"company_slug": "linear",
"total_board_jobs": 24,
"matched_count": 18,
"summary": { "remote_count": 18, "hybrid_count": 4, "onsite_count": 2, "with_salary_count": 14 },
"jobs": [
{
"title": "Senior Staff Backend Engineer",
"workplace_type": "remote",
"is_remote": true,
"seniority": "staff",
"salary": { "currency": "USD", "min": 210000, "max": 250000, "interval": "yearly" },
"tech_tags": ["TypeScript", "PostgreSQL", "GraphQL", "Node.js"]
}
]
}
```
### 4. `shopify_store_catalog` (`{"store_url": "allbirds.com", "limit": 10}`)
```json
{
"source": "shopify_public_json",
"status": "ok",
"store_url": "https://www.allbirds.com",
"returned_products": 10,
"collections_count": 25,
"products": [
{
"title": "Men's Wool Runners",
"price_min": 78.0,
"price_max": 98.0,
"compare_at_price_max": 110.0,
"max_discount_pct": 29.09,
"in_stock": true,
"available_variant_count": 6
}
]
}
```
### 5. `substack_public_posts` (`{"publication": "pragmaticengineer", "limit": 5}`)
```json
{
"source": "substack_public",
"status": "ok",
"publication_url": "https://newsletter.pragmaticengineer.com",
"free_count": 2,
"paywalled_preview_count": 3,
"posts": [
{
"title": "Inside Engineering Scaling",
"audience": "only_paid",
"is_paywalled": true,
"paywall_status": "paid_preview_only",
"body_markdown": "Public preview text only..."
}
]
}
```
### 6. `podcast_rss_transcripts` (`{"query_or_rss_url": "https://feeds.example.com/show.xml", "limit": 3}`)
```json
{
"source": "podcast_rss_podcasting20",
"status": "ok",
"show": {
"podcast_title": "Builders Weekly",
"owner_name": "Jane Host",
"owner_email": "host@buildersweekly.example",
"categories": ["Technology"]
},
"episodes_with_transcript": 3,
"episodes": [
{
"episode_title": "Ep. 42 — MCP in Production",
"transcript_status": "extracted",
"transcript": {
"format": "vtt",
"segment_count": 148,
"full_transcript": "Welcome back to Builders Weekly..."
}
}
]
}
```
---
## Need Scheduled Batch Runs / Cloud Datasets?
`mcp-public-intel` is designed for interactive, ad-hoc AI agent calls from your local machine.
When you need **scheduled cron runs, multi-company / multi-store batch extraction, webhooks, CSV/Parquet/JSON dataset exports, or managed cloud execution**, use our production **Apify Store Pay-Per-Event (PPE) Actors** (pay fractions of a cent per extracted record, zero infrastructure):
- **Apify Store Profile**: https://apify.com/pecopeko
- **Open-Source Actor Repo**: https://github.com/pecopeko/apify-ppe-actors
- `sec-edgar-8k-10k-item-extractor` — Batch SEC EDGAR 8-K/10-K/10-Q/Form D extraction with ASC 606 XBRL (`$0.004`/filing)
- `shopify-public-catalog-feed-extractor` — Multi-store Shopify catalog, SKU, price-drop & collection monitoring (`$0.001`/product)
- `substack-newsletter-posts-extractor` — Multi-publication Substack archive & RAG Markdown dataset pipeline (`$0.001`/post)
- `podcast-rss-transcript-extractor` — Bulk Apple Podcasts reviews, `<itunes:owner>` lead discovery & Podcasting 2.0 transcripts (`$0.002`/episode)
### Support Open-Source Maintenance
If `mcp-public-intel` saves you API fees or engineering hours, consider sponsoring ongoing maintenance via **[GitHub Sponsors](https://github.com/sponsors/pecopeko)**!
---
## Security & Fair-Access Guarantees
- **Strict SSRF Protection**: Blocks `localhost`, `.local`/`.internal` domains, cloud metadata IPs (`169.254.169.254`), RFC1918 private ranges, IPv6 loopback/link-local, IPv4-mapped private IPs, embedded URL credentials, and non-HTTP(S) schemes — and re-validates every HTTP redirect hop.
- **XXE / Billion-Laughs XML Guard**: Rejects `<!DOCTYPE` and `<!ENTITY` declarations in RSS feeds before XML parsing.
- **Polite Fair Access**: Identifies itself cleanly via `User-Agent`, rate-limits SEC EDGAR calls (`<= 8 req/sec`), and honors HTTP `429`/`503` `Retry-After` backoff.
- **Zero Bypass Policy**: Never bypasses CAPTCHAs, login walls, or paywalls. Paywalled Substack posts are transparently tagged `paywall_status: "paid_preview_only"`.
## License
MIT — see [LICENSE](LICENSE).
TDQS
Scored across 6 tools
Each tool targets a distinct data source and resource: SEC financials, SEC Form D filings, ATS job boards, Shopify catalogs, Substack posts, and podcast transcripts. The two SEC tools are cleanly separated by purpose (public company financials vs. private funding filings), so an agent can pick unambiguously.
All six tools follow a consistent snake_case, source-prefixed noun pattern (sec_*, ats_*, shopify_*, substack_*, podcast_*). The prefix conveys the underlying data source, which aids scanning.
Six tools, each covering a distinct public-intel source with no redundancy — a well-scoped set that fits comfortably in the 3-15 sweet spot. Nothing feels padded or missing at the count level.
The surface covers a solid cross-section of public data (financials, funding, jobs, e-commerce, newsletters, podcasts) with meaningful depth per source. Minor gaps remain — e.g. no cross-source entity resolution (ticker-to-startup) or general news/web search — but agents can work around these.