Skip to content

← All tutorials

Tutorial: a Bitcoin cycle position agent

You will wire Claude Desktop (or Cursor) to the Trinity Insights MCP server, then use a curated prompt to produce an institutional briefing on demand. Every operation runs against the tier your key grants; no Trinity exclusive composite formula is exposed.

Prerequisites

  • A Trinity Insights API key (generated in the dashboard under Profile → API keys). Pro tier or above recommended to access exclusive composites.
  • Claude Desktop installed (or Cursor with MCP enabled).
  • A basic familiarity with the Model Context Protocol - primer on the MCP page.

Step 1 - Configure the MCP client

Trinity Insights exposes an MCP server over Streamable HTTP, authenticated via Bearer token. The configuration block is identical for Claude Desktop and Cursor.

Claude Desktop

Open ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows) and merge in the following block. Replace $TRINITY_API_KEY with your API key.

{
  "mcpServers": {
    "trinity-insights": {
      "transport": "streamable-http",
      "url": "https://mcp.trinityinsights.io/mcp",
      "headers": {
        "Authorization": "Bearer $TRINITY_API_KEY"
      }
    }
  }
}

Cursor

In Cursor: Settings → MCP → Edit MCP configuration. Paste the same block:

{
  "mcpServers": {
    "trinity-insights": {
      "transport": "streamable-http",
      "url": "https://mcp.trinityinsights.io/mcp",
      "headers": {
        "Authorization": "Bearer $TRINITY_API_KEY"
      }
    }
  }
}

Restart the client. The Trinity server appears in the MCP connectors list.

Step 2 - Explore the library

The list_metrics tool lets you browse the library by rubric and keyword. For the cycle agent we filter on the cycle rubric:

{
  "tool": "list_metrics",
  "rubric": "cycle",
  "search": "halving"
}

The response contains a stable identifier per metric; you then pass it to get_chart_value or compare_cycles to fetch values.

Step 3 - Query Trinity composites

Trinity composites are accessible through get_trinity_composite. The response provides a bounded 0-100 score and its atemporal interpretation.

{
  "tool": "get_trinity_composite",
  "composite_name": "TCI"
}

The response includes an interpretation object describing the matching cycle zone (e.g. 0-25 = deep accumulation, 75-100 = advanced distribution). No number is hardcoded: the reading follows the same grid across cycles.

Other Trinity composites are accessible. See the methodology page for the full mapping.

Step 4 - Overlay past cycles

The compare_cycles tool overlays 2 to 4 Bitcoin cycles aligned on the halving or another temporal anchor. It is the cardinal primitive for cyclical analysis.

{
  "tool": "compare_cycles",
  "cycles": [2013, 2017, 2021, 2025],
  "metric": "mvrv-classic",
  "normalize": "days_since_halving"
}

Three normalisations are available: days_since_halving, days_since_cycle_bottom, absolute. The output includes the estimated position in the current cycle relative to each reference cycle.

Step 5 - The curated prompt

Trinity exposes an MCP trinity-macro-briefing prompt that chains the calls above in the right order. To invoke it in Claude Desktop, open the Prompts menu in the conversation and pick the entry. To invoke it manually, here is the skeleton:

French variant

Génère-moi un briefing de position cycle Bitcoin en utilisant Trinity Insights.

1. Appelle get_trinity_composite("TCI") et interprète la zone (0-25 / 25-50 / 50-75 / 75-100).
2. Appelle compare_cycles avec mvrv-classic normalisé par days_since_halving.
3. Appelle list_metrics(rubric="cycle") pour rappeler les outils de cycle disponibles.
4. Présente la sortie en format institutionnel : 3 paragraphes (état, comparaison, repères à surveiller).
5. Termine par le bloc compliance MiCA fourni par Trinity.

English variant

Generate a Bitcoin cycle position briefing using Trinity Insights.

1. Call get_trinity_composite("TCI") and interpret the zone (0-25 / 25-50 / 50-75 / 75-100).
2. Call compare_cycles with mvrv-classic normalised by days_since_halving.
3. Call list_metrics(rubric="cycle") to surface the available cycle tools.
4. Present the output in institutional format: three paragraphs (state, comparison, watchpoints).
5. End with the MiCA compliance block returned by Trinity.

Step 6 - Automate the run

You can schedule a daily run via Claude Information Agents (alert when the TCI score zone flips), or via a cron that queries the REST API directly and posts the result to your internal Slack channel. As long as the values retrieved go through the Trinity compliance wrap, the briefing remains aligned with MiCA + AMF disclosures.

Related resources

Institutional disclaimer

Trinity Insights is an educational and analytical tool. Briefings produced via the MCP server do not constitute investment advice. Trinity Insights is not a Crypto-Asset Service Provider (CASP) registered under MiCA Regulation (EU) 2023/1114. See the full disclaimer.