The Abu Dhabi Securities Exchange is opening a new route into its market data, allowing compatible artificial intelligence applications to retrieve official exchange information through natural-language requests rather than relying solely on conventional data terminals and individually built API connections.
ADX announced that its new service uses Model Context Protocol, or MCP, to connect AI applications including ChatGPT and Claude with approved exchange datasets. The exchange said the launch makes it the first securities exchange in the Middle East and North Africa to provide this type of direct access to live market data through major conversational AI platforms.
The development does not mean ADX market information automatically becomes part of an AI model’s built-in knowledge. Instead, the exchange operates an authenticated MCP endpoint that compatible applications can connect to. Once configured, an AI system can request information from ADX when a user asks a relevant question, with access determined by the user’s subscription, permissions and data entitlements.
ADX Adds a Conversational Layer to Market Data
At its simplest, the technology changes the interface between a user and the exchange’s information. Rather than navigating several screens or structuring a traditional API request, a connected user could ask an AI assistant a question in ordinary language and have the application retrieve the corresponding data from ADX. The exchange’s own examples include requests for recent prices and volumes, issuer disclosures, corporate actions and comparisons of historical sector performance.
The available information extends beyond headline share prices. The service can provide market-depth information for individual securities, bid-ask spreads, trading statistics and breakdowns of activity involving retail and institutional or foreign and local investors. Structured XBRL financial disclosures, index constituents and related reference information are also among the datasets described by ADX.
Abdulla Salem Alnuaimi, group chief executive officer of ADX, said: “AI is rapidly enabling how investors discover, analyse and act on information. As the first exchange in the MENA region to connect directly with major AI models, we are redefining how the financial community can take advantage of our market data to make informed decisions directly, intuitively, and efficiently. By removing technical barriers that once required a specific setup, everyone now has live access to ADX data through simple prompts. We believe this will contribute to greater market transparency, wider financial participation, and a more accessible investment ecosystem.”
How the MCP Connection Works
MCP is an open standard designed to let AI applications communicate with outside data sources, tools and systems through a common framework. Official MCP documentation describes the protocol as a way of connecting AI applications such as ChatGPT and Claude with external information and functions, instead of limiting their responses to information already available in the model’s context.
ADX has built its implementation around what it describes as a controlled, exchange-hosted endpoint. Users authenticate through an API key or an approved enterprise authentication method, while their permissions determine which datasets can be queried. ADX says the server can expose equities, derivatives, disclosures or combinations of datasets depending on the configuration. Information can also be delivered in structured formats including CSV, JSON and XBRL.
Compatibility with ChatGPT is also consistent with OpenAI’s current technical documentation. OpenAI says remote MCP servers can connect models to external data sources and can be used through supported ChatGPT and API workflows. The protocol therefore acts as the connection layer; it does not remove the exchange’s controls over authentication, licensing or the scope of information a particular user is authorized to retrieve.
Access Ranges From Individual Users to Institutions
ADX has divided the service into several access levels. Its current MCP plans begin with a free tier offering 100 tool calls per month and extend through paid packages with larger usage allowances and longer historical-data coverage. The highest listed self-service package costs about $13.61 (AED 49.99) per month and includes 4,000 tool calls and full historical depth.
Different market participants could use the same connection differently. A research team might query company disclosures or compare historical information, while a developer could incorporate ADX data into an internal analytical product. Brokers could build their own AI interfaces on top of the exchange feed, and individual investors could use natural-language questions to retrieve information available within their access level. Those applications would still sit on top of ADX’s data infrastructure rather than replacing it.
The launch also illustrates a broader change in how financial information can be distributed. Exchanges have traditionally supplied data through websites, terminals, vendor feeds and APIs. MCP introduces another layer in which an AI application can become the front end while the exchange remains the source behind the response.
