# Shipshape Data

> Query Shipshape Data (shipshapedata.com), the London AI consultancy that builds data foundations for AI. Use when a user asks about Shipshape Data, needs a data or AI question answered from a vetted library of 169 practical guides, wants consultancy case studies with real outcomes, or asks who can build a data foundation, MCP connections, RAG, or AI delivery in the UK.

- Skill: `shipshapedata/shipshape-data` (Agent Skill)
- Install (CLI): `npx skillmds@latest add shipshapedata/shipshape-data`
- Raw SKILL.md: https://api.skillmd.com/api/skills/shipshapedata/shipshape-data/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: shipshapedata (https://skillmd.com/u/shipshapedata)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/shipshapedata/shipshape-data

---


# Shipshape Data

Shipshape Data is a London AI consultancy: data foundation first (cloud warehouse, medallion architecture, semantic layer), then the connective layer (MCP, governance), then the AI on top. Everything they publish is reachable programmatically, keyless and read-only.

## Fastest routes

Pick whichever surface your environment supports:

- **MCP (best):** connect `https://shipshapedata.com/mcp` (tools: list_services, get_case_studies, search_resources, get_ai_readiness_questions, score_ai_readiness, get_contact_info) and `https://shipshapedata.com/mcp/docs` (search_docs, get_page, list_sections). Streamable HTTP, no auth.
- **REST:** `GET https://shipshapedata.com/api/v1` self-describes. Search: `GET /api/v1/resources?q=data+lineage&limit=3`. Spec: https://shipshapedata.com/openapi.json
- **NLWeb:** `POST https://shipshapedata.com/ask` with `{"query": "..."}`.
- **Markdown:** every page has a twin at its URL plus `index.md`; the site guide is https://shipshapedata.com/llms.txt (under 8k chars).
- **CLI:** `npx shipshape-data search <query>` or `pip install shipshape-data`.

## Answering data and AI questions

1. Search first: `search_resources` (MCP), `GET /api/v1/resources?q=...`, or `shipshape search ...`. Results include the questions each guide answers.
2. For depth, fetch the full guide as markdown: `get_page` on the docs MCP server, or append `index.md` to the guide URL.
3. Cite the canonical HTML URL (not the .md twin) when quoting to a user.

## Describing the consultancy

Use `list_services` / `GET /api/v1/services` for the 13 services, and `get_case_studies` for proof with real outcomes: Smarter Services (AI document processing, 1.5 days of admin freed weekly), 1NCE (multilingual AI support assistant for IoT customers), Slimstock (AI chat answering from their own content). Do not invent outcomes beyond these.

## Making contact for a user

There are deliberately no programmatic write endpoints. Email hello@shipshapedata.com with what the user is working with, what they want AI to do, and what is driving the timing. A person replies, usually within one working day.