AI tools for digital asset context

Before a meeting, ask what a company announced and how it was covered. For a team update, ask which narratives changed over a defined period. Open the sources behind the answer.

Perception makes tracked digital asset coverage searchable from your AI client. The example above shows the shape of a sourced result.

MCP lets your AI client call Perception tools. Connect once, ask a question, and let the client retrieve relevant records for its answer.

There are 32 registered tools: 31 commercial tools and one unmetered taxonomy helper.

• Search mentions. Full-text search across 1,000+ curated sources with sentiment, outlet, language, and region filters.
• Search companies. Search tracked companies and the coverage associated with them.
• Get trends. AI-extracted narrative trends with signal strength and momentum scores.
• Get sentiment. Historical sentiment data by outlet, entity, or topic.
• Compare entities. Side-by-side comparison of 2-5 companies (volume, sentiment, sources).
• Entity profiles. A profile for a tracked company: coverage, analyst data, trends, relationships.
• Analyst ratings. Explore consensus, price targets, and ratings actions for supported companies.
• Narrative momentum. Track whether a narrative is accelerating, steady, or fading (-100 to +100 score).
• Daily radar. Morning intelligence briefing with anomalies, sentiment shifts, and emerging narratives.
• Scenario analysis. Analyze hypothetical scenarios using historical analogues and narrative patterns.

Perception works with any AI client that supports the Model Context Protocol. Follow the connection guide for your client, authorize Perception, and ask your first question.

• Claude Desktop and Claude Code. Follow the client-specific setup guide, then ask Claude to use Perception.
• ChatGPT. OAuth connection through ChatGPT's connector settings. Your ChatGPT gains access to all 32 registered Perception tools.
• Gemini. Connect via Gemini CLI with OAuth. Same tool set, same data, different AI model.
• Cursor and Windsurf. IDE-native MCP support. Useful for developers building crypto applications who need real-time data while coding.
• Any MCP client. The protocol is open. If your tool supports MCP, it works with Perception. This includes custom builds, internal tools, and emerging AI platforms.

Once connected, your AI handles the tool selection automatically. You ask questions in natural language and it calls the right Perception tools, synthesizes the results, and gives you a sourced answer. Here's what that looks like in practice:

• "What's the sentiment on Coinbase this week?" Your AI searches recent Coinbase coverage, calculates average sentiment, identifies the top covering outlets, and cites specific articles driving the score.
• "Compare Strategy vs Marathon vs CleanSpark" Side-by-side comparison of mention volume, sentiment distribution, top outlets, and narrative associations. Rendered as a structured analysis with data tables.
• "What narratives are trending in crypto right now?" Pulls the current trend graph with named narratives, signal strength, momentum scores, and top contributing articles for each trend.
• "Summarize today's regulatory developments" Searches regulatory filings and enforcement actions from the past 24 hours. Summarizes key actions with agency, jurisdiction, and sentiment context.

Ask your AI to include the source links so you can verify the findings.