AlertoWatch

MCP FOR AGENTIC WORKFLOWS

Turn AI conversations into persistent crypto alert automation

Connect your AI agent to AlertoWatch through the Model Context Protocol. Let the agent reason about market context, create rule-based crypto alerts, choose notification workflows, and prepare follow-up actions — while AlertoWatch monitors the conditions continuously.

Your AI reasons. AlertoWatch keeps watching.

Explore the MCP tools
Streamable HTTPStatelessPublic API boundaryEntitlement-aware
AI Agent WorkspaceBTC reversal workflow
Agent connectedAlertoWatch MCP
BTC/USDTIllustrative workflow
DOWNTREND
EMA 13EMA 50Future crossover watch
EMA 13 below EMA 50Illustrative context
User
What is BTC trading at right now, and what does the current market context look like?
Agent

In this illustrative workflow, the connected AI agent classifies BTC’s current market structure as a downtrend.

AlertoWatch does not generate the analysis itself — it provides the persistent monitoring and automation layer.

User
Send me an alert as soon as the trend changes.
Agent

I’ll create an AlertoWatch alert that emails you when EMA 13 crosses above EMA 50.

To add confirmation, I’ll require price to be above EMA 50 when the crossover occurs.

Discord delivery is also available if you already have a channel configured. Should I include it?

User
No, email only. After the alert fires, create an alert for the opposite move.
Agent

No problem.

I’ll prepare the opposite EMA 13 crosses below EMA 50 alert now, keep it disabled, and use AlertoWatch After Trigger Actions to enable it when the reversal alert fires.

MCP tool calls

list_coinsResolving BTCUSDTDone
get_entitlementsChecking multi-condition and automation accessDone
create_alertPreparing opposite-direction alertDone
create_alertCreating reversal alert with email onlyDone
preview_notificationVerifying the email messageDone

Workflow status

BTC market resolved
Entitlements verified
Opposite alert prepared
Email-only policy selected
After Trigger Action armed
DONE

Your alert “BTC downtrend reversal” was created.

Primary rule
EMA 13 crosses above EMA 50
Confirmation
Price above EMA 50
Notification
Email only
Follow-up
Enable “BTC uptrend reversal” after trigger

The opposite rule is prepared in advance and enabled only after the primary alert event is accepted.

Deterministic illustrative sequence

Illustrative agent workflow — market and macro analysis come from the connected AI agent. AlertoWatch provides the persistent monitoring and automation layer.

FROM CONVERSATION TO AUTOMATION

Your AI can reason. AlertoWatch can keep monitoring.

An AI conversation is useful while the session is active. A rule-based alert remains useful after the conversation ends. The connected agent can translate your intent into supported AlertoWatch operations, while AlertoWatch stores the rule, monitors market data, records accepted trigger events, and sends the selected notifications.

The agent remains responsible for its analysis and recommendations. AlertoWatch remains responsible for deterministic alert validation, persistent monitoring, event handling, and supported automation.

Describe the outcome

Tell the agent what market event matters, which confirmation you want, and where the notification should go.

Create a valid rule

The agent uses AlertoWatch tools to resolve markets, check entitlements, and create or update supported alert conditions.

Keep monitoring

AlertoWatch evaluates the saved rule after the chat is over and records the trigger when the configured condition matches.

Continue the workflow

Supported After Trigger Actions can enable or disable prepared alerts and continue a controlled automation sequence.

MODEL CONTEXT PROTOCOL

A standard tool interface for AI clients

The Model Context Protocol gives AI clients a standard way to discover and call external tools. Through the AlertoWatch MCP server, a connected agent can work with supported alerts, watchers, coins, entitlements, event records, and notification actions without importing internal application code or accessing the database directly.

The AlertoWatch MCP server uses Streamable HTTP and operates statelessly. Authentication is provided per request with an AlertoWatch API token, and every operation still passes through the public API boundary, ownership checks, validation rules, and plan entitlements.

Endpoint
https://alertowatch.com/mcp
Transport
Streamable HTTP
Session model
Stateless
Authentication
Bearer API token
Capability
Tools
Internal database access
None

TOOL INVENTORY

Give your AI agent controlled access to AlertoWatch workflows

The MCP server exposes focused tools rather than unrestricted system access. The available operations cover discovery, alert and watcher management, event review, notification preview, and resend workflows.

Discover

list_coinsget_entitlementslist_alertsget_alertlist_watchersget_watcher

Resolve supported markets, inspect current account limits, and understand the alerts and watchers already owned by the authenticated user.

Create and manage

create_alertupdate_alertdelete_alertcreate_watcherupdate_watcherdelete_watcher

Create and maintain supported rule-based alerts and market-wide watchers through the same public API validation used by the application.

Control watchers

pause_watcherresume_watcher

Pause or resume a watcher explicitly without relying on ambiguous toggle behavior.

Review and continue

list_alert_eventslist_event_recordspreview_notificationresend_notification

Review accepted event records, preview the current rendered message, and resend supported notifications when needed.

WORKFLOW IDEAS

From one request to a repeatable monitoring workflow

Market analysis → alert

A connected agent interprets its own market data and creates a supported AlertoWatch rule for the condition the user wants monitored.

Alert → selected notification

Create an alert with email, Discord, Telegram, webhook, or in-app delivery according to the channels already configured and the alert’s current plan limits.

Watcher → market discovery

Create a Watcher Feed rule that scans one supported exchange universe for price change, volume change, or Bitcoin-relative performance conditions.

Trigger → prepared opposite alert

Prepare a disabled opposite-direction alert and use After Trigger Actions to enable it after the first alert fires.

Event → notification review

Inspect accepted event records, preview the current message, and resend a notification through supported configured channels.

Existing workflow → agent maintenance

Let the agent inspect, update, pause, resume, or delete the user-owned alerts and watchers that are within its authorized scope.

FOLLOW-UP AUTOMATION

Prepare the next action before the first alert fires

AlertoWatch After Trigger Actions allow an accepted trigger event to continue a controlled workflow. An alert can enable prepared alerts, disable selected alerts, or use a supported alert template action when the current plan and ownership rules allow it.

In the hero workflow, the agent prepares a disabled EMA 13 crosses below EMA 50 alert first. The primary EMA 13 crosses above EMA 50 alert then uses an After Trigger Action to enable the opposite alert after the reversal event is accepted.

The follow-up rule is prepared in advance. AlertoWatch does not ask an AI model to invent a new condition at trigger time.

Explore crypto alert automation

Controlled by design

  • Deterministic set operations
  • Same-user ownership validation
  • Normal alert validation
  • Entitlement enforcement
  • Event acceptance before side effects
  • Idempotent behavior

CONTROL AND BOUNDARIES

Agentic workflows without unrestricted system access

Authenticated per request

The client sends an AlertoWatch API token with each MCP request. Never place the token in page source, screenshots, logs, or public examples.

Stateless MCP server

The MCP transport does not keep private conversational memory between requests.

Public API boundary

Tools call supported AlertoWatch public APIs rather than internal service routes or database tables.

User-owned resources

Alerts, watchers, events, and follow-up references remain scoped to the authenticated user.

Validation and entitlements

MCP does not bypass DSL validation, ownership checks, active-alert limits, channel limits, or plan-specific capabilities.

No trade execution

AlertoWatch creates monitoring and notification workflows. It does not place orders or guarantee market outcomes.

GET CONNECTED

Connect a compatible AI client to the AlertoWatch MCP endpoint

Create an API token inside the authenticated AlertoWatch developer area, add the MCP endpoint to a compatible client, and pass the token securely in the Authorization header.

ClineCursorClaude Desktop through mcp-remoteStreamable HTTP MCP clients

API-token and write capabilities depend on the current AlertoWatch plan. Plan details and pricing are available after sign in.

Open developer setup

https://alertowatch.com/mcp

{
  "url": "https://alertowatch.com/mcp",
  "headers": {
    "Authorization": "Bearer api_<your-token>"
  }
}

Conceptual configuration — use the exact format required by your MCP client.

Frequently asked questions

What is the AlertoWatch MCP server?
The AlertoWatch MCP server exposes focused tools that compatible AI clients can use to work with supported alerts, watchers, coins, entitlements, event records, and notification actions through the public AlertoWatch API boundary.
Can an AI agent create crypto alerts for me?
Yes, when the authenticated account and current plan allow the requested operation. The agent can call supported alert tools, but the alert still passes through normal validation, ownership checks, and entitlement enforcement.
Does AlertoWatch provide the market or macro analysis?
No. Market interpretation and macro analysis come from the connected AI agent and its data sources. AlertoWatch provides persistent monitoring, alert evaluation, event handling, notifications, and supported follow-up automation.
Can the agent use email, Discord, Telegram, or webhooks?
An agent can create alerts with supported notification selections that are already configured and available to the user’s workflow. The selected channels remain subject to current account configuration and per-alert plan limits.
Can one alert prepare the next alert automatically?
Yes. A workflow can prepare another alert in advance and use supported After Trigger Actions to enable or disable it after the first alert event is accepted. The follow-up condition is not invented by an AI model at trigger time.
Does MCP give the agent direct database access?
No. The MCP server calls supported public AlertoWatch APIs. It does not expose direct database access or internal service routes.
Does MCP bypass AlertoWatch plan limits?
No. Tool calls remain subject to the current account’s entitlements, alert limits, notification limits, validation rules, and API permissions.
Which AI clients can connect to the AlertoWatch MCP server?
Currently documented setup examples include Cline, Cursor, Claude Desktop through mcp-remote, and compatible Streamable HTTP MCP clients. Exact configuration depends on the client.

Explore related AlertoWatch workflows

AI Crypto Alert BuilderCrypto AlertsCrypto Indicator AlertsMulti-Timeframe AlertsCrypto Alert Event HistoryEMA AlertsMACD Alerts

BUILD THE WORKFLOW

Let your AI agent create the rule. Let AlertoWatch monitor it.

Move from one-off market conversations to persistent, reviewable alert automation. Connect your AI client, create a supported workflow, and let AlertoWatch watch for the conditions after the chat ends.

Explore crypto alerts