Developer Hub

Developer Hub

Sentinel Watch exposes its observation platform through a live MCP (Model Context Protocol) server, enabling AI agents, chat interfaces, and automation tools to query, configure, and extend the platform programmatically. This page documents what is available today.

What Is MCP?

Model Context Protocol (MCP) is an open standard that allows AI models and agents to interact with external services through a structured tool interface. Instead of building custom API integrations for every AI product, MCP lets any compatible agent — Claude, GPT-based tools, custom LLM pipelines — connect to Sentinel Watch and invoke platform capabilities as tools.

Sentinel Watch operates a live MCP server hosted at app.sentinel-watch.org. Each enterprise client receives their own dedicated MCP connection with scoped credentials — giving their AI agents direct, isolated access to their observation programs, data, and configuration. No shared endpoints, no bespoke API code required.

Live Capabilities

The following capabilities are active on the Sentinel Watch MCP server today.

Platform Automation

Create, update, and manage observation programs, posts, and platform content through tool calls. Automate routine administrative tasks without manual intervention in the backend.

AI Agent Integration

Connect Claude, GPT-based agents, or any MCP-compatible AI to the Sentinel Watch platform. Agents can read platform state, trigger actions, and respond to events — enabling fully autonomous observation program management workflows.

Chat Agent Interface

Enterprise clients can interact with Sentinel Watch through a natural language chat interface powered by their MCP connection. Ask questions about observation programs, request status updates, or trigger configuration changes — all through conversation.

Workflow Automation

Build multi-step automated workflows that connect Sentinel Watch to your existing stack — CRMs, data warehouses, notification systems, or custom pipelines. Observation data can trigger downstream actions automatically on delivery.

Third-Party Integrations

Each client MCP connection supports integrations with external services including Discord, GitHub, email, and other platforms. Configure observation program notifications and reporting to flow directly into your preferred tools.

Custom Event Programs via API

AI clients can define and manage custom observation event taxonomies programmatically. Configure what observers watch for, adjust classification criteria, and retrieve structured results — all through the MCP tool interface.

Integration Architecture

The Sentinel Watch MCP server is hosted at app.sentinel-watch.org and exposes a structured set of tools that any MCP-compatible client can discover and invoke. Each enterprise client receives their own scoped connection. The integration pattern is straightforward:

  1. Connect — Point your MCP-compatible AI agent or client at your dedicated Sentinel Watch MCP endpoint. Authentication is handled via your provisioned connection credentials.
  2. Discover — The agent queries available tools. The server returns a structured manifest of all available actions, their parameters, and expected outputs.
  3. Invoke — The agent calls tools by name with the appropriate parameters. The server executes the action and returns structured results.
  4. Compose — Multiple tool calls can be chained within an agent’s reasoning loop to build complex, multi-step workflows without custom glue code.

Any enterprise client whose AI stack supports MCP — including Claude via Anthropic’s Claude.ai or Claude API, or any agent framework implementing the open MCP standard — can integrate with Sentinel Watch without bespoke development work on either side.

Compatible With

Claude (Anthropic)

Connect via Claude.ai’s MCP server settings or through the Anthropic API with MCP tool support. Claude can manage observation programs, query data, and handle configuration tasks conversationally.

Custom Agent Pipelines

Any agent framework that implements the MCP client specification can connect. This includes LangChain-based agents, custom LLM orchestration pipelines, and enterprise AI platforms with MCP support.

Automation Platforms

Non-AI workflow tools that support MCP or webhook-based triggers can also connect — invoking Sentinel Watch actions based on schedules, external events, or conditions in your own systems.

Roadmap

The following capabilities are in active development and will be added to the Developer Hub as they become available:

  • Public MCP endpoint documentation — Full tool manifest, parameter references, and authentication guide for enterprise integration partners
  • Observation data webhooks — Push structured observation results to your endpoint in real time on program completion or milestone events
  • Sandbox environment — A test MCP connection for integration partners to validate their agent’s tool calls before connecting to live program data
  • SDK and client libraries — JavaScript/TypeScript client library for direct MCP integration without a full agent framework
  • Observation stream API — Real-time event stream for clients who need live observation data as it is tagged and submitted by observers

Interested in Integrating?

Request Integration Access
Scroll to Top