The reference at /docs is a page a developer reads. It is not something a codegen tool or an AI coding assistant can consume directly, since there is no structured description of the endpoints behind the HTML. SimItNow now also publishes two machine-readable artifacts alongside it, and this article covers how to actually use them.
What /llms.txt and /openapi.json are
https://www.simitnow.com/llms.txt follows the llms.txt convention: a short, plain-text file naming the REST API base URL, the hosted-CSMS WebSocket base, the SIMITNOW-API-KEY auth header, and links to the full docs and the spec. It exists so an AI agent doesn't have to scrape a rendered HTML page just to learn where the API lives.
https://www.simitnow.com/openapi.json is a full OpenAPI 3.1 document: every current v1 and v2 endpoint, its request body fields, and its response shapes, in the same structured format that Postman, Insomnia, and every OpenAPI code generator already know how to read.
Prompting an AI coding assistant
An AI coding assistant that can fetch a URL (Claude, Cursor, Copilot Chat, and similar tools all support this) can read both files directly and write working code against them. A prompt like this one is copy-pasteable as written and produces a real, runnable client, because it only asks for exactly what the spec supports:
Read https://www.simitnow.com/llms.txt and https://www.simitnow.com/openapi.json,
then write a Python script using the requests library that:
1. Creates a charging station with id CP-001, pointed at ws://localhost:9000
using OCPP 1.6
2. Creates a virtual EV with id EV-01 and a 60 kWh battery
3. Connects EV-01 to CP-001's connector 1
4. Starts a transaction on that connector and prints the transaction id
Send my API key in the SIMITNOW-API-KEY header on every request.
That sequence, create a station, create an EV, connect it, start a transaction, is exactly the flow documented in the API reference's own "Full Session Lifecycle" guide, so an assistant reading the spec has everything it needs to get the field names, HTTP methods, and response shapes right on the first try.
A second, narrower prompt is just as useful when you only need an explanation rather than code: Read https://www.simitnow.com/openapi.json and explain what the /v2/csms endpoints do and how a chargepoint would connect to the result.
Pointing an API tool directly at the spec
Any tool that consumes OpenAPI documents can import https://www.simitnow.com/openapi.json the same way it would import a spec from your own project.
- Postman. Import → Link, paste the URL, and Postman builds a full collection with one request per endpoint, already typed with the request bodies the spec declares. Add your API key as a collection-level SIMITNOW-API-KEY header or environment variable so every request in the collection picks it up.
- Insomnia. Create → Import From URL, same spec URL. Insomnia generates one folder per resource (charging stations, EVs, the hosted CSMS instance) directly from the spec's tags.
- OpenAPI Generator. Point the CLI at the same URL to generate a typed client in almost any language, for example a Python package:
Swap -g python for typescript-axios, go, or any other generator target the tool supports.openapi-generator-cli generate -i https://www.simitnow.com/openapi.json -g python -o ./simitnow-client
When to read the human docs instead
The spec and the AI-assistant workflow above cover request/response shapes well, but a few things are still only written out in prose at /docs: the full list of probe diagnosis codes, the security profile options for the outbound CSMS connection, and the worked walkthrough of a complete session. Keep that page open alongside whatever client the assistant or the tool generates for you, since it is the page both artifacts here link back to.
Get your free API key and try one of the prompts above against your own SimItNow account.