Tools

The MCP server gives your agent 6 tools. Your agent picks the right one based on what you ask. You just talk naturally.

---

search_news

Hybrid semantic + keyword search across all indexed AI news. Returns results ranked by relevance with similarity scores.

When to use it: Your agent needs to make a recommendation — which model to use, which framework to pick, what hardware to buy — and you want it working with the latest data, not its training cutoff.

Example prompts:

• "What small models are good for coding and run on consumer hardware?"
• "Any recent papers on AI for drug discovery or genomics?"
• "Find news about GPU kernel optimization and inference efficiency"

Returns: Array of news events ranked by relevance. Each result includes two additional fields: similarity (relevance score, 0-1) and search_method ("semantic", "keyword", or "hybrid").

---

get_latest_news

The feed. Latest stories with optional filters for impact, entities, categories, models, and type.

When to use it: You want to browse what's new, optionally filtered down. The default for "what happened recently."

Example prompts:

• "Any new open-source models released this week?"
• "Show me high-impact news about NVIDIA and AMD"
• "What robotics research was published lately?"

Returns: Array of news events ordered by publish date (newest first).

---

get_news_detail

Full details on a specific news event, optionally with semantically related stories.

When to use it: You found an interesting event via search or the feed and want the full picture. Complete summary, sources, and related coverage.

Example prompts:

• "Tell me more about that NVIDIA Vera Rubin announcement"
• "Get full details on the reasoning model jailbreak paper"

Returns: Full news event object plus an array of related stories (if enabled).

---

get_weekly_digest

Weekly AI news digest with big-picture analysis, key developments, and AGI probability assessment.

When to use it: You want the big picture. What mattered this week, where the industry is heading, and how AGI probability is moving.

Example prompts:

• "Get this week's digest. What were the most important developments?"
• "How has the AGI probability changed over the last few weeks?"
• "What's the big picture in AI right now?"

Returns: Full digest with title, big-picture analysis, key developments, event counts, podcast info, and AGI probability data (current probability, delta, estimate in months, reasoning, category breakdown).

---

get_trending

Trending entities (companies/labs), categories, and model names with trend direction and change percentages.

When to use it: See which companies, categories, and models are getting the most coverage right now. The count reflects how many stories mention each entity, weighted by multi-source coverage. Compare against the prior period to spot what's gaining or losing attention.

Example prompts:

• "What's trending in AI this week?"
• "Which companies have been most active in the last month?"
• "Are there any emerging categories compared to last week?"

Returns: Three ranked lists: trending entities, categories, and models. Each with count, trend direction (up/down/new/stable), and change percentage vs. the prior period.

---

get_entities

Full directory of tracked AI entities. Companies, labs, universities, government bodies, and people.

When to use it: You need to discover valid entity slugs for filtering, or you want to see what the system tracks.

Example prompts:

• "What AI companies does Hyperterminal track?"
• "Show me all universities in the database"
• "Which government entities are being monitored?"

Returns: Array of entities with slug, display name, type, and aliases.