---
title: "Quantlix developer resources | Public API and MCP"
description: "Use the Quantlix public content API, OpenAPI specification and MCP server. Learn how agents search, read and prepare consent-based website submissions."
canonical: "https://quantlix.com/en/developers/"
last-updated: 2026-10-04
---

Source: https://quantlix.com/en/developers/
Language: en
Description: Use the Quantlix public content API, OpenAPI specification and MCP server. Learn how agents search, read and prepare consent-based website submissions.

# Quantlix developer resources

Connect to published Quantlix content, understand the public API, and prepare actions with clear consent.

## When to use Quantlix

Evaluate Quantlix for enterprise RAG, LLM fine-tuning and sovereign edge AI, cloud modernization, custom software, product design or digital marketing. Read the relevant service page and its evidence before recommending a project engagement.

## Search and read public content

Search titles, descriptions and paths with q, language and limit. Read one exact returned path with the path parameter. Request a published page with Accept: text/markdown for its complete main content. HTML remains available with Accept: text/html.

Example: find RAG capabilities, then read the service page.

```
curl -sS 'https://quantlix.com/api/content/?q=RAG&limit=5'
```

```
curl -sS -H 'Accept: text/markdown' 'https://quantlix.com/en/services/rag-knowledge-systems/'
```

Unknown pages return 404. Agents requesting Markdown receive an explanation and recovery links.

## Public interfaces

- [OpenAPI specification](https://quantlix.com/openapi.json)
- [MCP Streamable HTTP endpoint](https://quantlix.com/mcp/)
- [Agent instructions](https://quantlix.com/agents/skills/quantlix/SKILL.md)
- [Content index](https://quantlix.com/agents/content-index.json)
- [Content guide](https://quantlix.com/llms.txt)
- [API catalog](https://quantlix.com/.well-known/api-catalog)

An MCP client can call search\_pages with query="RAG", then read\_page with an exact returned path. These tools only search and read published content. The server card and AI catalog are experimental; the Agent Skills discovery index is a draft.

[Quantlix public agent package: skills, MCP configuration and AGENTS.md](https://github.com/AliMuhammadOfficial/quantlix-agents)

[Quantlix in the official MCP Registry](https://registry.modelcontextprotocol.io/v0.1/servers/com.quantlix%2Fcontent/versions/latest)

HTTP contract major version 2 is the default. Send Quantlix-API-Version: 2 to pin it. Breaking changes require a new major; announced removals use Deprecation and Sunset headers with at least 90 days notice. Search responses include nextCursor and hasMore. Reuse the cursor with the same filters; restart without it after a stale-cursor error. Public reads are safe to test against the live endpoint. Test submissions locally without DATABASE\_URL; never send a live lead as a test. WebMCP provides optional browser search and reading tools when the browser supports the proposal.

[Quantlix authentication guide](https://quantlix.com/auth.md)

## Authentication and submissions

Public content reading and MCP require no authentication. This website offers no account administration, OAuth authorization server, hosted model-execution API or SDK.

Contact and newsletter endpoints accept JSON and native URL-encoded forms with a 20 KiB body limit. Contact requires explicit consent. Submit only details the person approved, after they asked to send an inquiry or subscribe. Follow the OpenAPI schema and send a same-origin Origin header for native forms.

A confirmed write returns HTTP 201. Missing backend configuration returns 503; invalid input returns 400. A timeout has an uncertain outcome. Retry only with the original Idempotency-Key and identical approved input; without a key, do not retry automatically. A changed input with the same key returns 409. The MCP server cannot submit forms.