For developers

One API for laws and regulations.

Build your product on laws and regulations without setting up an AI pipeline yourself. Quonai searches laws, standards and your own frameworks, and gives answers you can trace to the article and paragraph.

Request access Visit developers.quonai.com

What you get

Everything an integration needs

You build the product and keep control of users and conversations. Quonai supplies the provisions, and answers that say what they rest on.

Search the source

Hybrid search, reranked by the language model, within the frameworks of your project. You get the provisions themselves back, not a summary.

Chat with sources

Answers with the provisions they rest on: code, title and the start of their text. Streaming, synchronous or by polling.

Scope analysis

Describe an organisation or a product. Quonai gives each framework a relevance from 0 to 1, with one sentence on why.

A fixed engine version

The engine's behaviour has a version, now 2026.09. Pinned per client or request, a new version only applies once you switch.

Projects and boundaries

One project per integration, with its own frameworks. The project sits in the credential: search and chat do not look beyond the frameworks of that project.

Your own framework

Internal standards or handbooks beside the law, kept apart from the shared corpus. We load them for you.

Model choice

Open models on our own hardware by default. Per client a hosted model from Anthropic, OpenAI or Google is possible; the question then goes there.

Developer portal

Documentation with Try it, quickstarts in TypeScript, Python and Java, a Postman collection, a knowledge base and a docs assistant.

Management, logs and MCP

Sandbox and production, secrets rotated without downtime, usage per call and webhooks. Also as an MCP server.

A first call

Token, question, answer

Two calls with curl. The same in TypeScript, Python and Java is on developers.quonai.com.

Shell

# 1. An access token (OAuth 2.0 client credentials)
TOKEN=$(curl -s -X POST https://api.quonai.com/v1/oauth/token \
  -u "$QUONAI_CLIENT_ID:$QUONAI_CLIENT_SECRET" \
  -d grant_type=client_credentials | jq -r .access_token)

# 2. A question, answered in one response
curl -s -X POST https://api.quonai.com/v1/chat \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"mode": "sync",
       "frameworkScope": ["NIS2"],
       "messages": [{"role": "user",
                     "content": "When must a significant incident be reported?"}]}' \
  | jq '{status, answer, citations: [.citations[] | .regulation + " " + .code]}'

Response (abridged)

{
  "status": "done",
  "answer": "…",
  "citations": ["NIS2 NIS2.ChIV.23.4"]
}

Getting started

Live in three steps

From a first conversation to a production client, with the same calls in the sandbox and in production.

1
Intro, 30 minutes
Your hardest question on the table. Together we choose the laws and standards that count for your product.
2
A sandbox in an afternoon
Your developers get a sandbox and ask the first questions from your own software the same afternoon.
3
Live in production
Ready to go live? Then you get a production environment, with speed and usage for every question.

During the preview the standard models are free; the sandbox has a limit of 2 million tokens a month.

Start in the sandbox

Request access. In a first conversation we set up your environment on developers.quonai.com.

Request access developers.quonai.com