hyperterminal/docs

Resources & Data Models

Auto-loaded MCP context and JSON schema reference

Overview

Unlike tools (which your agent actively decides to call), Resources provide passive context. Your MCP client loads these automatically in the background, giving your agent a baseline awareness of current AI events before you even ask your first question.

Available Resources

latest-news

URI: hyperterminal://news/latest

The 10 most recent AI news stories. Equivalent to calling get_latest_news with default parameters. (MIME type: application/json)

latest-digest

URI: hyperterminal://digest/latest

The most recently published weekly digest with big-picture analysis, AGI probabilities, and key developments. (MIME type: application/json)

How clients handle this: In Claude Desktop, resources appear in the context panel automatically. Other clients might load them eagerly or on demand. You don't need to configure anything. If your client supports resources, your agent is already reading them.

Resource Registry (At a Glance)

Prop

Type


Data Models

Every payload returned by Hyperterminal's tools and resources follows a strict schema. Here is the shape of the data your agent receives.

News Event

In this example, highlighted lines are the fields most agents rely on for filtering and relevance (impact_level, news_type, categories, entities, and model names).

News Event
{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "title": "OpenAI releases GPT-5 with native multimodal reasoning",
  "summary": "OpenAI announced GPT-5, featuring...",
  "main_url": "https://...",
  "source_account": "OpenAI (@OpenAI)",
  "impact_level": "HIGH", 
  "news_type": "news", 
  "categories": ["model-release"], 
  "entity_slugs": ["openai"], 
  "model_names": ["GPT-5"], 
  "published_at": "2026-02-27T12:00:00Z"
}

Prop

Type

Weekly Digest

In this example, highlighted lines are the AGI-tracker fields that carry the weekly forecast signal (current probability, delta, reasoning, and category evidence).

Weekly Digest
{
  "id": "...",
  "week_key": "2026-W09",
  "week_start": "2026-02-23T00:00:00Z",
  "week_end": "2026-03-01T00:00:00Z",
  "title": "AI Hardware Race Heats Up",
  "big_picture": "This week saw major moves in...",
  "key_developments": [...],
  "event_count": 47,
  "high_impact_count": 8,
  "podcast_status": "ready",
  "podcast_url": "https://...",
  "agi_probability": 12.5, 
  "agi_probability_delta": 0.3, 
  "agi_estimate_months": 36, 
  "agi_reasoning": "Continued progress in...", 
  "agi_category_breakdown": [ 
    { "category": "reasoning", "score": 15, "summary": "Notable progress in..." }, 
    { "category": "benchmarks", "score": 12, "summary": "New evaluation results..." }, 
    { "category": "cost_efficiency", "score": 10, "summary": "Training costs..." }, 
    { "category": "multimodal", "score": 8, "summary": "Cross-modal capabilities..." }, 
    { "category": "agents", "score": 14, "summary": "Autonomous agent frameworks..." }, 
    { "category": "scale", "score": 11, "summary": "Compute scaling trends..." } 
  ] 
}

Prop

Type


Reference Values

When filtering news or reading payloads, you will encounter these standardized enums and slugs.

Impact & Types

Impact LevelMeaning
HIGHMajor industry event: model launches, acquisitions, policy shifts
MEDIUMNotable development: research papers, partnerships, product updates
LOWMinor update: incremental changes, smaller announcements
News TypeMeaning
newsVerified, confirmed information
rumorUnverified, flagged as such in the feed

Entities & Categories

Entity Types:

  • company: OpenAI, Anthropic, Google DeepMind
  • lab: Meta AI, xAI
  • university: Stanford, MIT
  • government: EU Commission, White House
  • person: Sam Altman, Dario Amodei

Common Categories: llm · research · product · tools · api-platform · partnership · funding · infrastructure · hardware · safety · security · privacy · regulation · open-source · computer-vision · robotics

On this page