Patent Space

Tools (29)

• tool_adversarial_strategy

  Compare two firms' patent portfolios and generate strategic scenarios. Performs game-theoretic analysis of two firms' technology territories, identifying attack targets, defense priorities, and preemption opportunities. Args: firm_a: First company name (any language) or ticker. firm_b: Second company name (any language) or ticker. year: Analysis year (default: 2024). scenario_count: Number of strategic scenarios to generate (default: 3). Returns: Overview (overlap, tech di

• tool_applicant_network

  Build co-applicant network for a target applicant. Args: applicant: Company name (any language) or stock ticker. depth: Traversal depth from center node. min_co_patents: Minimum shared patents to keep an edge. Returns: Center entity and graph-style nodes/edges of co-filing links.

• tool_bayesian_scenario

  Bayesian patent investment simulation with data-driven priors. Three modes: "init" builds priors from patent data, "update" incorporates user's private information via Bayesian updating, "simulate" runs Monte Carlo simulation. All prior parameters are evidenced by patent/GDELT data. Args: mode: "init" (start scenario), "update" (adjust parameter), "simulate" (run). technology: Cluster ID or tech description (required for init). firm_query: Optional firm name for context. invest

• tool_citation_network

  Build a patent citation network around a patent or firm's top patents. Two modes: patent mode (BFS from a single patent) or firm mode (finds firm's most-cited patents, builds network from those). Returns graph structure with nodes, edges, and hub patent identification. Args: publication_number: Seed patent for patent mode. firm_query: Company name/ticker for firm mode. depth: BFS traversal depth (1 or 2). direction: "forward", "backward", or "both". max_nodes: Maximum netwo

• tool_cross_border_similarity

  Detect similar patent filings across international jurisdictions. Finds patents in target countries that match a firm's portfolio, a specific patent, or a technology description. Useful for FTO monitoring, licensing opportunities, and portfolio surveillance. Always includes legal disclaimer. Args: query: Firm name, publication number, or technology description. query_type: "firm", "patent", or "text" (default: "firm"). target_jurisdictions: Country codes to search (default: ["CN","

• tool_cross_domain_discovery

  Discover cross-domain technology clusters related to a query. Finds technology clusters in different CPC sections that share embedding-space proximity with the source domain. Args: query: CPC code (e.g. "H01M") or free-text technology description. top_n: Maximum number of results (default: 10). exclude_same_domain: Exclude clusters from the same CPC section (default: True). min_similarity: Minimum cosine similarity threshold (default: 0.3). Returns: Source info, list of cr

• tool_entity_resolve

  Resolve a company name to its canonical form. Accepts names in any language/format. Uses 3-level matching: exact, normalized, and fuzzy. Args: query: Company name to resolve (any language). limit: Maximum results to return. Returns: List of matching entities with confidence scores.

• tool_firm_patent_portfolio

  Get a firm's patent portfolio analysis. Accepts firm name in any language (Japanese/English) or stock ticker. Returns patent counts, technology distribution (CPC), filing trends, and co-applicant relationships. Args: firm: Company name (any language) or stock ticker (e.g., "7203"). date: Cut-off date in YYYY-MM-DD format (default: all time). include_expired: Include expired patents in analysis. Returns: Dict with entity info, patent count, CPC distribution, filing trends,

• tool_firm_tech_vector

  Get a firm's precomputed technology vector and diversity metadata. Fast lookup from pre-computed table. Returns patent_count, dominant_cpc, tech_diversity (entropy), tech_concentration, and the full tech_vector. Data available for ~4,300 firms, years 2015-2024. Args: firm_query: Company name (any language), stock ticker, or firm_id. year: Analysis year (default: 2024). Falls back to latest available year. Returns: Dict with firm_id, year, patent_count, dominant_cpc, tech_diversity

• tool_gdelt_company_events

  Fetch GDELT events/GKG and cached five-axis features for a firm. Returns pre-cached GDELT media signals: tone, event_count, theme_diversity, geographic_spread, and source_diversity. Data covers 2020Q1-2024Q4 for ~46 major firms. Args: firm_query: Company name (any language) or stock ticker. date_from: Start date as YYYYMMDD integer (e.g., 20200101). date_to: End date as YYYYMMDD integer (e.g., 20241231). Returns: Dict with five-axis features, event timeline, and summary.

• tool_invention_intelligence

  Analyze a technology description for prior art, FTO risk, and whitespace. Given a natural language description of an invention or technology, identifies the relevant patent cluster, finds prior art, assesses freedom-to-operate risk, and discovers whitespace opportunities. Args: text: Technology description in natural language (Japanese or English). max_prior_art: Maximum number of prior art results (default: 20). include_fto: Include FTO risk assessment (default: True). include

• tool_ma_target

  Recommend M&A acquisition targets based on patent portfolio analysis. Three strategies: "tech_gap" finds firms with complementary technology (fills acquirer's weak areas), "consolidation" finds firms with high overlap (strengthens market position), "diversification" finds firms in unrelated CPC sections (opens new markets). Args: acquirer: Acquiring company name or ticker. strategy: "tech_gap", "consolidation", or "diversification". top_n: Number of targets to return (default: 10).

• tool_patent_compare

  Compare multiple firms by patent volume, CPC mix, and trend. Args: firms: List of company names or tickers. cpc_prefix: Optional CPC prefix filter. date_from: Start date in YYYY-MM-DD format. date_to: End date in YYYY-MM-DD format. Returns: Per-firm portfolio metrics plus shared and unique CPC classes.

• tool_patent_detail

  Get detailed information for a specific patent. Args: publication_number: Patent publication number (e.g., "JP-2020123456-A"). JP granted patents use kind code B1: "JP-7637366-B1". JP applications use kind code A: "JP-2020-123456-A". Returns: Full patent record including CPC codes, assignees, inventors, and citations.

• tool_patent_market_fusion

  Combine patent portfolio strength with market signals for strategic analysis. Produces a fusion score combining tech_strength, growth_potential, diversity, and market_sentiment (GDELT). Supports investment screening, M&A target identification, and license partner matching. Args: query: Company name, technology description, CPC code, or patent publication number (e.g., JP-7637366-B1). query_type: "firm", "technology", "text", or "patent". Auto-detected if None. purpose: Analysis pur

• tool_patent_search

  Search patents by keyword, CPC code, applicant, or date range. Scans the full patent table (13M+ JP patents). May be slow under heavy I/O load. For technology-area analysis, prefer tech_landscape (by CPC) or startability_ranking (by firm or cluster) which use pre-computed data. Args: query: Free-text search (Japanese or English). Searches title and abstract. cpc_codes: Filter by CPC classification codes (e.g., ["G06N3"]). applicant: Filter by applicant/assignee name (partial match)

• tool_patent_valuation

  Score patent, portfolio, or technology area value with royalty rate reference. Three modes: "firm" evaluates company portfolio strength, "patent" scores a single patent, "technology" evaluates a tech area by CPC code or keyword. All modes provide industry-specific royalty rate benchmarks. Args: query: Company name, patent number, or CPC code / technology keyword. query_type: "firm" (portfolio), "patent" (single), or "technology" (tech area). purpose: "licensing", "portfolio_ranking

• tool_portfolio_evolution

  Track how a firm's technology portfolio evolved over time. Shows year-by-year changes in dominant technology areas, diversity, and concentration. Identifies emerging and declining technology investments. Summarizes strategic shifts. Args: firm_query: Company name (any language) or stock ticker. year_from: Start year (default: 2016). year_to: End year (default: 2024). Returns: Timeline with yearly metrics, emerging/declining clusters, and strategic shift summary.

• tool_sales_prospect

  Identify and rank patent licensing sales targets. Finds firms that need your technology, generates "why they need it" narratives, and provides approach guides with deal structure recommendations. No contact information is included. Args: firm_query: Licensor company name (your firm). patent_or_tech: Patent number, technology description, or cluster_id. query_type: "patent", "text", or "cluster" (default: "cluster"). target_count: Number of prospects to return (default: 10). Re

• tool_similar_firms

  Discover firms with similar patent portfolios. Computes cosine similarity between firm technology vectors across ~4,300 companies. Useful for M&A candidate discovery, partnership identification, and competitive benchmarking. Args: firm_query: Company name (any language) or stock ticker. top_n: Number of similar firms to return (default: 10). year: Analysis year (default: 2024). Returns: Ranked list of similar firms with similarity scores, shared clusters, and unique techno

• tool_startability

  Compute startability score for a firm-technology pair. Startability measures a firm's readiness to enter or compete in a technology area, combining tech_fit (phi_tech), organizational capability, and dynamic trajectory. Uses pre-computed surface data for fast lookup. Args: firm_query: Company name (any language), stock ticker, or firm_id. tech_query_or_cluster_id: Cluster ID (e.g., "H01M_0", "G06N_0") or CPC code. Use tech_clusters_list to browse available clusters. year: A

• tool_startability_delta

  Compute change in startability scores over time. Shows which firm-technology pairs gained or lost the most startability. Requires multi-year startability surface data. Args: mode: "by_firm" (cluster deltas for one firm) or "by_tech" (firm deltas for one cluster). query: Firm name/ticker (by_firm) or cluster_id/CPC code (by_tech). year_from: Start year (default: 2020). year_to: End year (default: 2024). top_n: Number of results (default: 20). direction: "gainers" (highes

• tool_startability_ranking

  Rank tech clusters or firms using precomputed startability surface. Two modes: "by_firm" ranks technology clusters by startability for a given firm (what tech areas can this firm enter?). "by_tech" ranks firms by startability for a given cluster (who can enter this tech?). Args: mode: "by_firm" (rank clusters for one firm) or "by_tech" (rank firms for one cluster). query: Firm name/ticker (by_firm) or cluster_id like "H01M_0" (by_tech). year: Analysis year (default: 2024). Availabl

• tool_tech_clusters_list

  List technology clusters with optional CPC filtering. Returns 607 technology clusters derived from CPC classification. Each cluster has a label, patent_count, top_applicants (top firms), and top_terms (keywords). Use cpc_filter to narrow by CPC section (e.g., "H01" for electrical elements, "G06" for computing). Args: sort_by: Sort field — "patent_count" (default) or "cluster_id". top_n: Maximum clusters to return (default: 200, max: 2000). cpc_filter: CPC prefix filter (e.g., "H01"

• tool_tech_fit

  Compute phi_tech fit components for a firm and technology cluster. Measures how well a firm's patent portfolio aligns with a technology cluster. Returns CPC overlap ratio, citation proximity, co-inventor score, and combined phi_tech score. Args: firm_query: Company name (any language), stock ticker, or firm_id. tech_query_or_cluster_id: Cluster ID (e.g., "H01M_0", "G06N_0") or CPC code (e.g., "H01M"). Use tech_clusters_list to browse available clusters. year: Analysis year

• tool_tech_gap

  Analyze technology gap and synergy between two firms. Quantifies complementarity, overlap, and acquisition fit using startability surface data. Useful for M&A due diligence, partnership evaluation, and competitive positioning. Args: firm_a: First company name or ticker. firm_b: Second company name or ticker. year: Analysis year (default: 2024). Returns: Overlap areas, each firm's strengths, synergy score, overlap score, and acquisition fit classification.

• tool_tech_landscape

  Analyze filing trends and top applicants in a technology area. Args: cpc_prefix: CPC prefix filter (e.g., "G06N"). query: Optional free-text filter on title/abstract. date_from: Start date in YYYY-MM-DD format. date_to: End date in YYYY-MM-DD format. granularity: Time granularity, "year" or "quarter". Returns: CPC trend, top applicants, growth areas, and total patent count.

• tool_tech_trend

  Analyze time-series technology trends with growth rates and new entrants. Given a technology query, returns year-by-year filing trends, growth rates, acceleration, new entrant detection, and sub-area breakdown. Supports Japanese/English keywords, CPC codes, and cluster IDs. Args: query: Technology keyword, CPC code, or cluster_id. cpc_prefix: Optional CPC prefix filter. year_from: Start year (default: 2016). year_to: End year (default: 2024). top_n: Maximum results per cate

• tool_tech_trend_alert

  Detect hot and cooling technology trends with market signals. Automatically identifies rapidly growing technology clusters, firms entering multiple hot areas, and declining clusters. Integrates GDELT media signals where available. All signals are expressed probabilistically. Args: year_from: Start year for trend analysis (default: 2020). year_to: End year (default: 2024). min_growth: Minimum growth rate to flag as "hot" (default: 0.3). top_n: Maximum results per category (defau