Build with AI agents

Everything on this site is also published in machine-readable form, so an AI assistant or coding agent can read the actual integration contract instead of guessing it. Point your tool at the files below.

Files

File	URL (UAT)	What it contains
`llms.txt`	docs.keyspace-qat.com/llms.txt	The index (llmstxt.org format): key facts — base URLs, authentication, response envelope, rules — and a link to every document below
`llms-full.txt`	docs.keyspace-qat.com/llms-full.txt	The complete API reference as one Markdown file — the same content as the Full API reference page (about 80 KB, roughly 20k tokens)
`openapi.json`	docs.keyspace-qat.com/openapi.json	The OpenAPI 3.0.3 specification: every endpoint, schema and error code. Its default server is UAT
Guides as Markdown	onboarding.md · guides/meeting.md · guides/members.md	Each page of this site as raw Markdown

The getting-started page, each guide and the full reference also have a Copy Markdown button and an Open in menu that hands the page's Markdown to ChatGPT or Claude.

Environments

Every environment serves the same set of files for the API version it runs. The API hosts also serve llms.txt and llms-full.txt at their root, where tools that follow the llmstxt.org convention look for them.

Environment	Docs host	API host
UAT	`https://docs.keyspace-qat.com`	`https://api.keyspace-qat.com`
Development	`https://docs.keyspace-dev.com`	`https://api.keyspace-dev.com`

Read the files of the environment you integrate against — credentials are not shared across environments. The **Version** line at the top of llms-full.txt and info.version in openapi.json tell you which API version they describe; fetch them again when that changes instead of keeping an old copy.

Use with an AI assistant

Claude, ChatGPT, Gemini

Give the assistant the URL and tell it to stay inside the document:

Read https://docs.keyspace-qat.com/llms-full.txt. Using only that document,
write a TypeScript function that gets an access token and lists the meeting
rooms free to book tomorrow from 10:00 to 11:00.

"Using only that document" keeps the model from inventing endpoints or fields. If the assistant cannot browse, download llms-full.txt and attach it.

Claude Code

Add a pointer to your repository's CLAUDE.md, so the agent fetches the contract before it writes integration code:

## Keyspace API
Before writing Keyspace integration code, fetch
https://docs.keyspace-qat.com/llms.txt and follow its links to the guide the
task needs. The OpenAPI spec is https://docs.keyspace-qat.com/openapi.json.

Cursor

In chat, type @Docs → Add new doc, paste https://docs.keyspace-qat.com/llms-full.txt, then mention @Docs in your prompts.

Your own agent or script

curl -fsSL https://docs.keyspace-qat.com/llms.txt
curl -fsSL https://docs.keyspace-qat.com/llms-full.txt -o keyspace-llms-full.txt

import requests

docs = requests.get("https://docs.keyspace-qat.com/llms-full.txt", timeout=30).text

Generate a client from the OpenAPI spec

npx @openapitools/openapi-generator-cli generate \
  -i https://docs.keyspace-qat.com/openapi.json \
  -g typescript-fetch \
  -o ./keyspace-client

In Postman: Import → Link → https://docs.keyspace-qat.com/openapi.json. The generated requests use the spec's default server, UAT (https://api.keyspace-qat.com). To browse the same spec interactively, open the OpenAPI reference.

Which file should I use?

You want to…	Use
let an agent find the right document by itself	`llms.txt`
give a model the whole contract in one prompt	`llms-full.txt`
generate a typed client, mocks or request validation	`openapi.json`
work on one module only (rooms & desks, or members)	that guide's `.md`
read it yourself	these pages

Keep secrets out of prompts

Caution

Never paste credentials or personal data into an AI tool. Do not put your app_secret, access or refresh tokens, or real member data — names, emails, phone numbers, card numbers, face photos — into a chat, a prompt file or an agent's instructions. Give the agent placeholders and load credentials from environment variables at runtime.

On this page