Market dashboards are designed primarily for humans. They may show trends, levels, signals, and risk information clearly on screen, but an AI agent cannot necessarily understand or interact with that information in a structured way.
BTC Sentinel explores a different model: what if the same market intelligence could be understood by both a human and an AI agent?
WebMCP made it possible to turn the website itself into an agent-accessible interface without replacing the human experience.
What it does
BTC Sentinel is an agent-ready Bitcoin market intelligence monitor.
The dashboard gives a human user an immediate view of:
- BTCUSD market state
- Multi-timeframe structure across H1, M30, and M5
- Support and resistance levels
- Setup status
- Risk context
- What needs to happen next before a setup becomes actionable
At the same time, BTC Sentinel exposes structured read-only WebMCP tools that allow a compatible AI agent to access and reason over the same intelligence.
The current WebMCP tools include:
get_market_stateget_market_structureget_key_levelsevaluate_setupexplain_signal
For example, a user can ask ChatGPT for the current BTC market state or support and resistance levels. The agent can retrieve the structured information and explain why BTC Sentinel is currently showing WAIT rather than forcing an entry.
How we built it
BTC Sentinel is a lightweight web application built with HTML, CSS, and JavaScript.
WebMCP tools are registered directly from the webpage using the browser's model context API. Each tool exposes a focused part of the application's market intelligence so an agent can request only the information needed for the task.
The human-facing dashboard and the agent-facing tools use the same underlying market snapshot. This keeps the visual interface and agent responses aligned.
The public challenge build is intentionally read-only and uses a static Bitcoin market snapshot. This makes the demo deterministic and reliable for judging without exposing broker credentials or enabling trade execution.
Challenges we ran into
The biggest challenge was designing the application for two different consumers at the same time: a human looking at a visual dashboard and an AI agent requesting structured context.
We also had to make the WebMCP interface useful rather than simply exposing the entire page as unstructured data. This led us to separate market state, structure, key levels, setup evaluation, and signal explanation into focused tools.
Testing browser support and WebMCP discovery across environments was another important challenge.
Accomplishments that we're proud of
We built a working public WebMCP application where the same market intelligence can be inspected by both a person and an AI agent.
During testing, ChatGPT was able to work with BTC Sentinel's market information, retrieve multi-timeframe support and resistance levels, and reason that BTCUSD being close to M5 resistance supports a WAIT decision rather than chasing a BUY.
The result demonstrates a practical human-agent workflow instead of treating AI as a separate chatbot beside the application.
What we learned
The most important lesson was that agent-ready websites benefit from structured capabilities, not just more text.
A good WebMCP tool should have a clear purpose and return enough context for the agent to reason while keeping the human interface understandable and useful.
What's next for BTC Sentinel
The next step is to connect the WebMCP layer to safely refreshed market data while preserving the read-only intelligence model.
Future versions could expand market coverage, improve explainability, add richer scenario evaluation, and allow humans and agents to investigate changing market conditions together in real time.
BTC Sentinel's goal is simple: make market intelligence understandable to humans and directly usable by AI agents through the web.
Log in or sign up for Devpost to join the conversation.