---
title: "Pack Studio for AI Agents: MCP Server and API"
url: "https://www.pack-studio.com/agents"
description: "Connect an AI assistant to Pack Studio's MCP server or REST API to search the site, read pages as Markdown, look up sourced figures and estimate savings."
updated: "2026-11-19"
type: "WebPage"
---

For AI agents

# Pack Studio for AI Agents

**Built to be read by machines.**

Pack Studio publishes a Model Context Protocol (MCP) server, a REST API and a Markdown copy of every page, so assistants and crawlers get the same sourced answers people do.

[Server card](https://www.pack-studio.com/mcp/server-card) · [llms.txt](https://www.pack-studio.com/llms.txt)

## Connect in one line.

Add this URL as a custom connector or remote MCP server in any assistant that supports MCP: `https://www.pack-studio.com/mcp`

- **Protocol:** Streamable HTTP, revision 2026-07-28, with 2025-11-25, 2025-06-18 and 2025-03-26 supported for older clients.
- **Authentication:** none. Every tool is read-only except request_demo, which acts only when the person asks to be contacted.
- **Discovery:** a server card at `/mcp/server-card` and an AI catalog at `/.well-known/ai-catalog.json`.

## What an agent can do.

**Pack Studio MCP tools**

| Tool | What it does | Effect |
|---|---|---|
| search_pack_studio | Searches features, answers, the User Guide, articles, industries and the glossary | Read-only |
| get_page | Returns any pack-studio.com page as Markdown | Read-only |
| get_facts | Looks up published figures with their condition, method and source | Read-only |
| list_features | Lists every feature with its status, summary and page | Read-only |
| estimate_engineering_savings | Estimates engineering hours and labor saved per launch, with the Lucid Motors benchmark | Read-only |
| request_demo | Asks the Pack Studio team to contact a person who requested it | Creates a lead, only after the person confirms |

## Every machine-readable surface.

| Surface | URL | What it is for |
|---|---|---|
| MCP server | https://www.pack-studio.com/mcp | Model Context Protocol over Streamable HTTP, protocol 2026-07-28 and earlier. Tools: search, page reader, sourced facts, savings estimate, feature list, demo request. |
| Server card | https://www.pack-studio.com/mcp/server-card | Static description of the MCP server for discovery before connecting. |
| AI catalog | https://www.pack-studio.com/.well-known/ai-catalog.json | Domain-level list of the MCP server, the API, llms.txt and the facts dataset. |
| OpenAPI | https://www.pack-studio.com/openapi.json | REST endpoints: search, savings estimate, lead and part intake. |
| API catalog | https://www.pack-studio.com/.well-known/api-catalog | RFC 9727 linkset pointing to the OpenAPI description. |
| Search API | https://www.pack-studio.com/api/search?q=returnable+racks | Ranked results across features, answers, the guide, articles and the glossary. |
| Markdown twins | https://www.pack-studio.com/features.md | Every page as Markdown at the same URL with Accept: text/markdown, or with .md added. |
| llms.txt | https://www.pack-studio.com/llms.txt | Curated index of the site for language models; llms-full.txt has every page in one file. |
| Facts registry | https://www.pack-studio.com/data/claims.json | Every published figure with its condition, method and source. Also as CSV. |
| Knowledge graph | https://www.pack-studio.com/data/graph.json | Merged schema.org JSON-LD for every entity on the site. |
| Sitemaps | https://www.pack-studio.com/sitemap.xml | Sitemap index with page, feature, answer, article, glossary and video sitemaps. |
| Feeds | https://www.pack-studio.com/feed.xml | Articles and news as RSS; the changelog has its own feed at /changelog.xml. |

## Try it.

- Read a page as Markdown: `curl -H "Accept: text/markdown" https://www.pack-studio.com/lucid`
- Search: `curl "https://www.pack-studio.com/api/search?q=returnable+racks"`
- Estimate savings: `curl "https://www.pack-studio.com/api/roi?n=1500&before=15&after=5&rate=70"`
- Every figure with its source: `curl https://www.pack-studio.com/data/claims.json`

## Crawler and AI policy.

Search engines, AI search and AI assistants are welcome. robots.txt allows all crawlers and declares Content-Signal search=yes, ai-input=yes, ai-train=yes. Any figure an agent cites should come from the facts registry, which lists the source, condition and method for every number.

[robots.txt](https://www.pack-studio.com/robots.txt) · [Facts & Figures](https://www.pack-studio.com/facts) · [Knowledge graph](https://www.pack-studio.com/data/graph.json)

## Agent questions.

### Do agents need an API key?

No. Search, pages, facts and estimates are open. The demo request tool creates a lead, so call it only when the person asks to be contacted and has given their own details.

### What should an agent cite?

The page URL, and for any figure, the source listed for it in the facts registry at /data/claims.json.

### Which assistants can connect?

Any client that supports the Model Context Protocol over Streamable HTTP, current revision 2026-07-28 or the earlier revisions back to 2025-03-26.

## Building something with Pack Studio data?

Tell us what you are building and what you need from the API.

[Contact us](https://www.pack-studio.com/contact)
