BrunoSan Global AI News

Not a news feed.

От сообществаРаботаетБез входаГлобальныйБесплатноТолько чтение

Что умеет

  • News Search: Full-text search across the current live AI and technology news corpus. Searches titles and summaries of all stored articles. Best for: finding recent developments on a topic, researching
  • News Get Entity: All articles and statistics for a company, product, person, or region. Returns mention count, recent articles, entity type, and co-occurring entities (what else appears alongside this
  • News Trending: Top topic clusters for a given day — what is moving the AI world right now? Returns grouped topic bundles ranked by burst_score (with article_count as tiebreaker), each with a represent

Какие данные видит

Нужен ли аккаунт

Не нужен: сервер работает без входа

Not a news feed. A signal verification system.

131,000+ AI & tech articles, ~1,000 added every day, drawn from 1,832 canonical source objects across 141 monitored domains. UUID-native, EU-hosted, GDPR-compliant.

Every article carries 7 deterministic relevance scores — AI, business, Mittelstand, regulation, risk, cost, operational — computed without an LLM. Reproducible, auditable, stable across runs.

21,000+ canonical entities. Not name variants: canonical objects with a deterministic UUID, alias resolution and type normalisation. "Anthropic", "Anthropic PBC" and "anthropic's" resolve to one object, not three.

19,000+ clusters built by deterministic heuristics, with burst score, novelty score and co-occurrence networks. Ask what moved today, who reported it first, how fast it spread, and whether the signal looks organic or manufactured.

9 read-only tools. Every answer traceable to objects, sources, relations and timestamps.

Built by DoWell UG, Hamburg — part of BrunoSan Intelligence, covering AI News, Adtech, ArXiv, Biotech, Crypto, Cyber, Finance, Geopolitics, Quantum, Regulatory and Robotics, connected by the LENS cross-vertical layer.

Список инструментов сервера (9)

Технические названия из tools/list. Нужны только разработчикам.

news_searchFull-text search across the current live AI and technology news corpus. Searches titles and summaries of all stored articles. Best for: finding recent developments on a topic, researching a company, product or trend, discovering coverage of specific events. Example queries: - "Latest news about Anthropic's new model" - "Find all articles mentioning the EU AI Act from the last 7 days" - "Search for every article where Sam Altman and regulation appear together — ranked by recency" Args: params (SearchInput): - query (str): Search term (any language) - limit (int): Number of results, default 10 - offset (int): Pagination offset, default 0 Returns: str: JSON with fields: - total_found (int): Total matches in database - count (int): Number returned in this response - offset (int): Current offset - has_more (bool): More results available - next_offset (int|null): Offset for next page - articles (list): Articles with id, title, teaser, source, url, published
news_get_entityAll articles and statistics for a company, product, person, or region. Returns mention count, recent articles, entity type, and co-occurring entities (what else appears alongside this entity in the news). Best for: researching OpenAI, Google, a CEO, a product like ChatGPT, or understanding what a company is being mentioned with. Example queries: - "Show me everything about Mistral AI" - "Who is Jensen Huang most often mentioned with in AI news?" - "Full entity profile for DeepSeek — articles, mention count, and who they co-occur with most" Args: params (EntityInput): - entity (str): Entity name (e.g. 'OpenAI', 'Elon Musk', 'ChatGPT') - limit (int): Number of articles to return, default 10 Returns: str: JSON with fields: - entity (str): Matched entity name - entity_type (str): Type: company/person/product/region - total_mentions (int): Total mentions across all articles - articles (list): Most recent articles mentioning this entity - co_occurring_entities (list): Frequently co-mentioned entities with count
news_trendingTop topic clusters for a given day — what is moving the AI world right now? Returns grouped topic bundles ranked by burst_score (with article_count as tiebreaker), each with a representative title, top entities, and counts. Clusters are formed by the BrunoSan news pipeline from daily article intake. Best for: daily briefings, 'what happened in AI today?', trend detection. Example queries: - "What are today's trending AI topics?" - "What were the top AI clusters on March 15th, 2026?" - "Today's top 5 trending clusters — dominant entities and event types, not just headlines" Args: params (TrendingInput): - date (str): Date YYYY-MM-DD, empty = latest available day - limit (int): Number of clusters to return, default 10 Returns: str: JSON with fields: - date (str): Date used for query - total_clusters (int): Available clusters for this day - clusters (list): Each with cluster_id, label, topic, event_type, article_count, burst_score, representative_title, source, url, top_entities
news_daily_digestComplete daily digest: all clusters, top articles, and statistics for one day. The full overview of a day's AI news — ideal for morning briefings, newsletter preparation, or 'summarize today's AI news'. Also includes the URL of the published human-readable news page. Example queries: - "Give me today's complete AI news digest" - "Summarize the AI news from last Monday for my weekly board update" - "Pull the full digest for 2026-03-20 — clusters, source count, and top articles per cluster for a client briefing" Args: params (DigestInput): - date (str): Date YYYY-MM-DD, empty = latest available day Returns: str: JSON with fields: - date (str): Date of the digest - published_url (str): URL of the published page on brunosan.de - stats (dict): articles_crawled, clusters_formed, sources_active - top_clusters (list): Top 5 clusters, each with label, topic, article_count, headline, top_articles (title, source, url)
news_get_sourcesStatistics for canonical source objects, historical article domains and feed operations. Shows which domains contribute the most articles, total database stats, and daily averages. Use this to understand the data foundation behind the news answers, or to verify source diversity and coverage. Best for: 'what sources does this data come from?', transparency checks. Example queries: - "Which German AI news sources do you cover?" - "What are the top 10 most active sources by article volume?" - "Which sources cover AI news in Japanese or Chinese — volume and daily averages for Asian market coverage assessment" Args: params (SourcesInput): - limit (int): Number of top sources to return, default 20 Returns: str: JSON with fields: - db_stats (dict): total_articles, unique_sources, total_entities, unique_entities, daily_avg_articles - top_sources (list): Sources with domain and article_count
news_search_entitiesSearch across the current entity index: companies, people, products, and regions. Shows which entities are known in the database and how often they are mentioned. Useful for discovery before using news_get_entity. Best for: 'which CEOs are most mentioned?', 'which German companies appear in AI news?', 'is [company X] in the database?' Example queries: - "Is 'DeepMind' in your entity database?" - "Find all AI companies from Germany in your entity database" - "Which AI CEOs are mentioned most in the last 30 days — rank by mention count, filter to type: person" Args: params (EntitySearchInput): - query (str): Search term within entity names - type (str): Optional filter: 'company', 'person', 'product', or 'region' - limit (int): Number of results, default 20 Returns: str: JSON with fields: - query (str): Search term used - type_filter (str|null): Type filter applied - count (int): Number of entities found - entities (list): Entities with name, type, mention_count
news_signal_strength7 deterministic signal scores per article: AI relevance, business relevance, Mittelstand relevance, regulation, risk, cost, operational — all 0-100, computed without LLM. Reproducible. Auditable. UUID-stable. Filters articles by a chosen score_type and minimum value, with optional quality flags (exclude duplicates, low-signal, non-AI). All scores read from article_fact_profile using the currently activated pipeline version. Example queries: - "Show me today's highest-signal AI regulation articles" - "Find all articles with risk_relevance above 80 today — no duplicates, no low-signal noise" - "Which articles have both high regulation_relevance AND high business_relevance? I need the intersection — regulatory AND commercially significant" - "Top 10 cost_relevance articles from this week — building a CFO briefing on AI infrastructure spend signals" - "Score today's AI news by operational_relevance for a CTO who needs to know what's breaking in production AI systems right now" Args: params (SignalStrengthInput): - score_type: which of the 7 scores drives the ranking - min_score: minimum value for that score (0-100) - since_days: lookback window (1-30 days, default 1) - exclude_duplicates / exclude_low_signal / exclude_non_ai: quality filters - limit: number of articles to return Returns: str: JSON with fields: - score_type (str): selected ranking score - min_score (int): applied threshold - since_days (int): applied lookback window - filters_applied (dict): which quality flags were filtered - count (int): number of articles returned - articles (list): each with id, title, teaser, source, url, published, and a `scores` block carrying all 7 deterministic scores plus quality flags
news_decision_maker_briefAI news clusters ranked by a transparent composite formula: composite = burst_score * (tier1_source_count + 1) * (hard_numbers_count + 1) * ai_relevance_score Role-aware via cluster scope filtering (CDO/CTO/CEO/Compliance). The composite formula itself is identical across roles — what changes is which clusters are eligible, so the formula stays auditable and reproducible. The (+1) smoothing prevents zeroing-out for fresh clusters that have not yet picked up tier-1 coverage or hard data. Role scopes: - cdo: all clusters - cto: primary_event_type in (product_launch, benchmark_result, security_incident, infrastructure, model_release) - ceo: business_relevance_score >= 50 - compliance: compliance_signal_count > 0 OR risk_signal_count > 0 Example queries: - "What does a CDO need to know in AI today?" (role=cdo) - "Generate a Monday morning AI briefing for our board — CEO perspective, top 5 items only" - "What's the highest-priority AI news for a compliance officer? Clusters with official sources and compliance signals, ranked by urgency" - "I'm a CTO preparing for a board meeting. Top 3 most decision-relevant AI developments this week — with hard numbers, not just headlines" - "Compare today's priority rankings for role=ceo vs role=cto — where do they diverge? What matters to one but not the other?" Args: params (DecisionMakerInput): - role: cdo | cto | ceo | compliance - date: YYYY-MM-DD or empty for latest - limit: number of clusters to return Returns: str: JSON with date, role, composite_formula, role_filter_applied, count, and clusters list (each with composite_score and all input components for transparent verification).
news_risk_radarActive risk and compliance signals ranked by combined_risk_score: combined_risk_score = (risk_signal_count * 2) + (compliance_signal_count * 3) + (official_source_count * 1) Open formula. Auditable. The AI News equivalent of the Regulatory MCP's reg_score — covering all AI news, not just regulatory documents. Example queries: - "What AI risk and compliance signals are active today?" - "Show all clusters with compliance_signals above 2 — what's legally relevant in AI right now?" - "Daily risk scan for our AI governance team — top 5 clusters by combined_risk_score, with source breakdown and official source count" - "Which AI risk signals came from official sources like governments or regulators? Filter official_sources > 0, rank by compliance signal density" - "Map today's risk radar against last week's — which new risk clusters appeared? Which escalated? Delta report for weekly risk committee" Args: params (RiskRadarInput): - date: YYYY-MM-DD, empty for latest - min_official_sources: filter to clusters with at least N official sources - min_compliance_signals: filter to clusters with at least N compliance signals - limit: number of clusters to return Returns: str: JSON with date, formula, filters_applied, count, and clusters list (each with combined_risk_score and the three component counts).