Aiera

Retrieve all available watchlists with their IDs, names, and descriptions. Used to find valid watchlist IDs for filtering other tools.

get_available_watchlists

Get detailed information about specific company documents including summary, content, and metadata. RESPONSE SIZE WARNING: This tool returns full document content which can be extensive, especially for annual reports and detailed press releases. WHEN TO USE: - Use this when you need complete document content and summary - Use this when you need full metadata for specific documents - For browsing documents by type/date, use find_company_docs instead WORKFLOW: Use find_company_docs first to obtain valid company_doc_ids.

get_company_doc

Retrieve document categories for filtering company documents. Used to find valid category values for find_company_docs. Requires at least one company-scoping filter: bloomberg_ticker, isin, ric, permid, sector_id, or subsector_id. The returned categories are the ones that exist for documents within that scope. Global (unscoped) aggregation is not supported.

get_company_doc_categories

Retrieve keywords for filtering company documents. Used to find valid keyword values for find_company_docs. Requires at least one company-scoping filter: bloomberg_ticker, isin, ric, permid, sector_id, or subsector_id. The returned keywords are the ones that exist for documents within that scope. Global (unscoped) aggregation is not supported.

get_company_doc_keywords

Retrieve core instructions that define how to use Aiera tools and data effectively. CRITICAL — CALL THIS TOOL FIRST BEFORE ANY OTHER TOOLS ARE CALLED: - You MUST call this tool at the beginning of each session to retrieve baseline instructions for working with Aiera data, and guidance on tool selection, data interpretation, and response composition.

get_core_instructions

Get the CURRENT analyst rating and price target for one or more companies — THE tool for "what is [provider]'s current rating / price target on [company]?", for any entitled provider. Answers from the provider's own frequently-updated coverage data where available and automatically falls back to the provider's newest covering research document (sector/industry notes included) otherwise — no manual document workflow needed. RETURNS: per requested identifier, the matching entries: company name, security identifiers, current rating, price target with currency, and the date the values last changed. Each entry carries a ``source``: - ``"coverage_feed"``: live provider coverage data; attribute in text with the per-provider ``as_of`` timestamp (e.g. "per Barclays coverage as of Aug 4"). There is no citable document — do not fabricate a citation link. - ``"document"``: extracted from the provider's newest covering note (``document_id`` / ``document_title`` identify it, and a ``citation_information`` block is included — cite it inline as with any research doc). ALWAYS state the note's ``published_date`` with the value — it is as-of that note, not live. Identifiers with no match from any queried provider are listed under ``unmatched``. WHEN TO USE: - Any "current rating / price target" question, regardless of provider. - Batch companies into ONE call (up to 50 identifiers) instead of calling per company. - Omit provider_ids for every provider with a view; pass it when the user names firms. - Use get_research_metadata_ratings instead when the question is about a SPECIFIC report ("what rating is in this note?"). WORKFLOW: get_current_ratings(identifiers=[...]) -> answer. The manual document workflow (find_research -> get_research_metadata_ratings) is only for unmatched identifiers, historical/prior values, or reading the underlying report.

get_current_ratings

Retrieve detailed summary(s) about one or more equities, filtered by bloomberg_ticker (comma-separated string). Summaries will include past and upcoming events, information about company leadership, recent financials, and within which indices the equity is included. For multiple companies, pass bloomberg_ticker as a single comma-separated string (e.g. "AAPL:US,MSFT:US") in one call.

get_equity_summaries

Get detailed information about a specific event including the full transcript. RESPONSE SIZE WARNING: This tool returns large text content. The transcript field can contain thousands of words for earnings calls and presentations. Consider using search_transcripts for targeted content extraction instead of reading full transcripts. WHEN TO USE: - Use this when you need the complete transcript text - Use this when you need full event metadata and summary - For finding specific quotes or topics, prefer search_transcripts instead LIMITATIONS: - Transcripts are not available for future events - This tool accepts one event_id per call. For multiple events, issue the calls in parallel where possible rather than sequentially — each call is independent. For broad keyword search across many events, use search_transcripts instead. WORKFLOW: Use find_events first to obtain valid event_ids.

get_event

Get detailed information about a specific SEC filing including summary and content. RESPONSE SIZE WARNING: This tool returns large text content. SEC filings (especially 10-K and 10-Q) can contain extensive text. Consider using search_filings for targeted content extraction instead of reading full filing content. WHEN TO USE: - Use this when you need the complete filing summary and metadata - Use this when you need full filing content for a specific document - For finding specific disclosures or risk factors, prefer search_filings instead WORKFLOW: Use find_filings first to obtain valid filing_ids.

get_filing

Retrieve financial statement data for a company. Use this tool to get detailed financial metrics for a specific company and fiscal period. Available statement types: - Income Statement: Revenue, costs, operating income, net income, EPS - Balance Sheet: Assets, liabilities, equity, cash position, debt - Cash Flow Statement: Operating cash flow, CapEx, free cash flow, financing activities WORKFLOW: Use find_equities first to verify the bloomberg_ticker if needed.

get_financials

Retrieve a grammar template that contains output formatting instructions. Grammar templates define how you should structure, format, and present responses — including tone, style, and common language patterns. CRITICAL — CALL THIS TOOL BEFORE PRODUCING A RESPONSE: - You MUST call this tool at least once per session with `template_type='general'` to retrieve baseline formatting instructions before generating any response that uses Aiera data. TEMPLATE TYPES: - 'general': Baseline formatting instructions applicable to all Aiera data responses (default)

get_grammar_template

Get all equities that are members of a specific stock market index (e.g. S&P 500). WORKFLOW: call get_available_indexes first to resolve the index name into a valid index id or short_name, then pass it here. Do not guess index ids. Results are paginated (25 per page) — page through to retrieve full membership for large indices.

get_index_constituents

Retrieve KPIs (Key Performance Indicators) and business segment data for a company. Use this tool to get company-specific operational metrics and segment breakdowns, including: - KPIs: Subscriber counts, unit sales, average selling prices, same-store sales, etc. - Segments: Revenue and metrics broken down by business unit, geography, or product line These metrics are company-specific and vary by industry. Not all companies report segment data. WORKFLOW: Use find_equities first to verify the bloomberg_ticker if needed.

get_kpis_and_segments

Retrieve financial ratios for a company. Use this tool to get ratio metrics for a specific company and fiscal period, including: - Profitability ratios (ROE, ROA, profit margins) - Liquidity ratios (current ratio, quick ratio) - Valuation ratios (P/E, P/B, EV/EBITDA) - Leverage ratios (debt-to-equity, interest coverage) WORKFLOW: Use find_equities first to verify the bloomberg_ticker if needed.

get_ratios

Get detailed information about a specific research report including summary and content. RESPONSE SIZE WARNING: This tool returns full research content which can be extensive. Consider using search_research for targeted content extraction instead of reading full reports. WHEN TO USE: - Use this when you need the complete research report content and summary - Use this when you need full metadata for a specific research report - For finding specific topics across research, prefer search_research instead WORKFLOW: Use find_research first to obtain valid document_ids. EXTRACTION STATUS: The response includes ``prose_available`` (bool) and ``content_length`` (int) fields per document. Some research notes are chart-only or table-only and extract no meaningful prose — in that case ``prose_available`` is ``false`` and ``content_length`` is small. When ``prose_available`` is false, do NOT attempt to summarize the document's content; the title/abstract/authors are still trustworthy for identification purposes, but the body itself is empty.

get_research

Retrieve all available research asset classes with their names and document counts. Used to find valid asset class values for filtering research tools. WHEN TO USE: - Use this to discover available asset classes (e.g. "Equity", "Fixed Income") before filtering research - Use this to understand the distribution of research across asset classes WORKFLOW: Use this tool to obtain asset class names, then pass them to find_research or search_research.

get_research_asset_classes

Retrieve all available research asset types with their names and document counts. Used to find valid asset type values for filtering research tools. WHEN TO USE: - Use this to discover available asset types (e.g. "Common Stock", "Corporate Bond") before filtering research - Use this to understand the distribution of research across asset types WORKFLOW: Use this tool to obtain asset type names, then pass them to find_research or search_research.

get_research_asset_types

Search for research authors by name or provider. Returns author IDs and display names. Used to find valid author_ids for filtering find_research and search_research tools. WHEN TO USE: - Use this to look up author IDs before filtering research by a specific analyst or team - Use this to discover authors associated with a specific research provider ALWAYS CALL THIS when the user mentions a specific analyst, team, or group (e.g., "economics team", "equity strategy", "Stan Shipley"). Pass `search` with the name and `provider_id` if known; then pass the resolved ID(s) as `author_ids` to find_research or search_research. TEAM NAME NORMALIZATION: Team names in the system often differ from how users refer to them colloquially. For example, the Evercore ISI economics team is stored as "EVRISI EcoTeam". This makes the lookup step essential — do not guess or pattern-match team names without resolving them here first. Team-authored publications are frequently published as standalone documents that would be missed if filtered only by provider. WORKFLOW: Use this tool to obtain author_ids, then pass them to find_research or search_research.

get_research_authors

Retrieve all available research country codes with their names and document counts. Used to find valid country code values for filtering research tools. WHEN TO USE: - Use this to discover available country codes before filtering research - Use this to understand the distribution of research across country codes WORKFLOW: Use this tool to obtain country code names, then pass them to find_research or search_research.

get_research_country_codes

Retrieve the complete source metadata record for a research document. RETURNS: structured metadata including publication status and dates, publisher organization and author roles, covered issuers and securities, subject/sector/region/country classifications, product series, and file details. WHEN TO USE: - Use for complete or esoteric metadata needs — classifications, product series, publisher/author roles, coverage lists, or publication details not present in find_research / get_research results. - For ONLY the analyst rating or price target, use get_research_metadata_ratings instead: it returns a compact rating/target payload (a few hundred bytes) rather than this full metadata document. - For a targeted slice, pass ``fields`` with dot-paths (call get_research_metadata_fields first to discover valid paths). WORKFLOW: find_research or search_research -> document_id -> get_research_metadata. Repeated elements beyond ``max_list_items`` are truncated with a ``<key>__truncated`` marker of shown/total counts.

get_research_metadata

List every available metadata field path for get_research_metadata, with per-field coverage and typical size share. RETURNS: a catalog of dot-paths (elements and @attributes), each with docs_pct (how often the field is present across documents) and mean_size_pct (its typical share of document bytes), plus the node names hidden by default and usage notes for the ``fields`` and ``max_list_items`` parameters. WHEN TO USE: call once before using get_research_metadata's ``fields`` parameter to choose paths. Prefer high-coverage paths; avoid or cap large ones (e.g. issuer lists) unless the user needs them.

get_research_metadata_fields

Get just the analyst ratings and price targets from a specific research report — a focused subset of get_research_metadata, and the fastest, most token-efficient way to answer "what is the rating / price target in this report?" RETURNS: per issuer and security: the Current (and Prior, when published) rating and target price with currency, security identifiers (RIC/Bloomberg/ISIN/CUSIP) for disambiguation, and rating/target-price actions when the publisher supplies them. Some publishers rate at the document level instead — that is returned as ``document_rating`` with its source noted. Sector reports may return many issuers, each with their own rating and target — match the requested company by name or security identifier in the response. The response includes a ``citation_information`` block for the report — cite it inline as with any research doc. WHEN TO USE: - Prefer this over get_research_metadata for ANY rating / price-target question — the response is a few hundred bytes instead of the full metadata document. - For other metadata (classifications, coverage lists, product series, publication details), use get_research_metadata instead. - An empty ``issuers`` list means the publisher did not include structured ratings in this document (common for macro/economics notes). - ``"target_price": null`` means the publisher omits target prices from this document's structured data — NOT that the report lacks one. Where possible the tool recovers the value from the document itself (returned with ``"source": "document_text"``); if it is still null, check the same document's full text via get_research before reporting the target as unavailable. Never substitute a target from an older document. WORKFLOW: find_research or search_research -> document_id -> get_research_metadata_ratings. For a company's CURRENT rating/target (not a specific report), use get_current_ratings instead — it answers directly without a manual document lookup.

get_research_metadata_ratings

Retrieve all available research product focus values with their names and document counts. Used to find valid product focus values for filtering research tools. WHEN TO USE: - Use this to discover available product focus values before filtering research - Use this to understand the distribution of research across product focuses WORKFLOW: Use this tool to obtain product focus names, then pass them to find_research or search_research.

get_research_product_focuses

Retrieve all available research providers with their IDs, names, and descriptions. Used to find valid provider IDs for filtering research tools. ALWAYS CALL THIS when the user mentions a specific research provider or bank (e.g., HSBC, Goldman Sachs, Morgan Stanley, BofA, Evercore). Pass the provider name as the `search` parameter to resolve it into an aiera_provider_id, then pass that ID as `aiera_provider_ids` to find_research or search_research. Omitting the provider filter when a specific provider is named returns results from all providers — that is almost never what the user wants. BANK DISAMBIGUATION: When a bank or financial institution is named, default to treating it as a research provider (search for reports authored BY the bank) rather than as a company being researched. Only treat the bank as the research subject if the user explicitly asks about the bank's own financials, earnings, or business operations.

get_research_providers

Retrieve all available research region types with their names and document counts. Used to find valid region type values for filtering research tools. WHEN TO USE: - Use this to discover available region types before filtering research - Use this to understand the distribution of research across region types WORKFLOW: Use this tool to obtain region type names, then pass them to find_research or search_research.

get_research_region_types

Retrieve all available research subjects with their names and document counts. Used to find valid subject values for filtering research tools. WHEN TO USE: - Use this to discover available research subjects before filtering research - Use this to understand the distribution of research across subjects WORKFLOW: Use this tool to obtain subject names, then pass them to find_research or search_research.

get_research_subjects

Get a sales commentary email in full: metadata, synopsis, referenced research and the email body as ordered sections. WORKFLOW: Use find_sales_commentary or search_sales_commentary first to obtain valid document_ids. EXTRACTION STATUS: The response includes ``prose_available`` (bool) and ``content_length`` (int). When ``prose_available`` is false the body extracted no meaningful prose — do NOT summarize the content; the subject, synopsis and authors are still trustworthy for identification.

get_sales_commentary

Retrieve all available sectors and subsectors with their IDs, names, and hierarchical relationships. Used to find valid sector/subsector IDs for filtering other tools.

get_sectors_and_subsectors

Get detailed information about a specific Third Bridge expert insight event. ABOUT THIRD BRIDGE: Third Bridge provides expert network insights - interviews and forums with industry experts, former executives, and specialists who provide unique perspectives on companies, industries, and market trends. RESPONSE SIZE WARNING: This tool returns full transcript content from expert interviews, which can be extensive. The response includes agenda, key insights, and complete transcripts. WHEN TO USE: - Use this when you need complete expert interview content - Use this for full context on expert insights and analysis - For finding specific topics across multiple events, use find_third_bridge_events with filtering LIMITATIONS: - If you need multiple events, make separate sequential calls (one event_id per call) WORKFLOW / WHICH ID TO PASS: - If the previous tool was ``find_third_bridge_events``: pass its ``event_id`` field as ``thirdbridge_event_id`` (e.g. ``thirdbridge_event_id="9c73e90f758aeb839236cc555b80de86"``). Note that ``find_third_bridge_events`` items ALSO contain ``aiera_event_id`` — both identifiers refer to the same event, but ``thirdbridge_event_id`` is preferred here because it's the provider's canonical id. - If the previous tool was ``search_thirdbridge``: pass its ``aiera_event_id`` field as ``aiera_event_id`` (integer). ``search_thirdbridge`` doesn't surface the Third-Bridge id. - Only one of the two is required. Passing both is redundant; passing neither is an error.

get_third_bridge_event

Get confirmed and estimated upcoming events within a date range. Requires one of the following: bloomberg_ticker (comma-separated string), watchlist_id, index_id, sector_id, or subsector_id. For multiple companies, pass bloomberg_ticker as a single comma-separated string (e.g. "AAPL:US,MSFT:US") in one call.

get_upcoming_events

Get all equities that are members of a specific watchlist. WORKFLOW: call get_available_watchlists first to resolve the watchlist into a valid watchlist_id, then pass it here. Do not guess watchlist ids. Results are paginated (25 per page) — page through to retrieve full membership.

get_watchlist_constituents

Simple ping test to verify server is running.

ping

Report research documents that informed your final answer. Call this tool exactly once after composing your response, passing the IDs of every research document whose content you used. Only report documents whose content meaningfully contributed to your answer; do not report documents that were retrieved and then discarded as off-topic, nor documents that only appeared in a listing without being read. Pass up to 100 research document IDs per call (the ``document_id`` values returned by ``find_research``, ``search_research``, or ``get_research``).

report_research_usage

Report sales commentary emails that informed your final answer. Call this tool exactly once after composing your response, passing the IDs of every email whose content you used. Only report emails whose content meaningfully contributed to your answer; do not report emails that were retrieved and then discarded as off-topic, nor emails that only appeared in a listing without being read. Pass up to 100 document IDs per call (the ``document_id`` values returned by ``find_sales_commentary``, ``search_sales_commentary`` or ``get_sales_commentary``).

report_sales_commentary_usage

Semantic search within company document chunks for specific content, disclosures, or topics. WHEN TO USE THIS TOOL: - Use this when you need to find specific content within company documents (presentations, press releases, investor materials, etc.) - Use find_company_docs FIRST to identify relevant documents by date/company/category, then use this tool to search within their content - Use this for targeted content extraction rather than reading full documents RETURNS: Relevant company document chunks with context, metadata, and relevance scores. Results are individual sections/chunks, not full documents. Use get_company_doc for complete document content. WORKFLOW EXAMPLE: 1. User asks: "What did Apple say about sustainability in their investor presentations?" 2. First call find_company_docs with company_id to get relevant document IDs 3. Then call search_company_docs with query_text='sustainability' and the company_doc_ids

search_company_docs

Semantic search within SEC filing content for specific topics, disclosures, or risk factors. WHEN TO USE THIS TOOL: - Use this when you need to find specific content (risk factors, disclosures, financial discussions) within SEC filings - Use find_filings FIRST to identify relevant filings by date/company/form type, then use this tool to search within their content - Use this for targeted content extraction rather than reading full filings RETURNS: Relevant filing chunks with context, filing metadata, and relevance scores. Results are individual sections/chunks, not full filings. Use get_filing for complete filing content. WORKFLOW EXAMPLES: - **Broad topical query** (e.g. "What are companies saying about supply chain risk in 10-Ks?"): call search_filings directly with query_text and form_number — the pre-filter via find_filings is unnecessary. - **Targeted single-company query** (e.g. "What are Tesla's main risk factors?"): optionally narrow first with find_filings(bloomberg_ticker='TSLA:US', form_number='10-K') → pass the resulting filing_ids into search_filings. The pre-filter is OPTIONAL; if scope is unclear, skip find_filings and call search_filings directly.

search_filings

Semantic search within research content for specific topics, analyses, or insights. WHEN TO USE THIS TOOL: - Use this for broad queries where many documents are expected — it returns relevant text chunks across the research corpus - Use this for targeted content extraction from research reports - For narrow queries (≤5 relevant documents), prefer find_research → get_research instead, which returns complete documents rather than chunks SCOPING TO A COMPANY: When the question is about a specific company, pass `bloomberg_ticker` (or `isin` / `ric` / `permid`). Without it the semantic search ranks chunks from the WHOLE research corpus, so a query like 'Intel 3Q25 preview' returns the closest-sounding passages about any company. The scope is resolved server-side and the matched companies are echoed back as `resolved_equities` so you can confirm the filter hit the intended company. RETURNS: Relevant research chunks with context, metadata, and relevance scores. Results are individual sections/chunks, not full research documents. ANALYST RATINGS: Result chunks may include analyst rating fields — `security_ratings_primary`/`_secondary` (security-level, e.g. "Overweight", "Equal Weight", "Outperform"), `issuer_ratings_primary`/`_secondary`, and `sector_industry_ratings_primary`/`_secondary`. Values are the provider's own rating labels, passed through verbatim, so compare within a provider, not across.

search_research

Semantic search within sales commentary emails (a provider's sales-desk notes) for what the desk is saying about a topic or company. WHEN TO USE THIS TOOL: - "What is the desk saying about ...", "how did sales react to [company]'s print", thematic questions across many emails - For a specific provider / company / week, prefer find_sales_commentary → get_sales_commentary, which return whole emails rather than sections SCOPING TO A COMPANY: pass `bloomberg_ticker` (or `isin` / `ric` / `permid`). Without it the search ranks sections from the WHOLE corpus. The scope is resolved server-side and the matched companies are echoed back as `resolved_equities`. RETURNS: relevant email sections with the parent email's metadata (document_id, title, series, provider, authors, date) and relevance scores. Results are sections, not whole emails; pass document_id to get_sales_commentary for the full email. Sales commentary reacts to research: rating and price-target actions it relays belong to the provider's research. Attribute them to the provider and prefer the referenced research document (or get_current_ratings) as the authoritative source.

search_sales_commentary

Semantic search within Third Bridge expert interview transcripts for specific topics, insights, or discussions. WHEN TO USE THIS TOOL: - Use this when you need to find specific content (expert opinions, industry insights, competitive analysis) within Third Bridge interviews - Use find_third_bridge_events FIRST to identify relevant interviews by date/company, then use this tool to search within their content - Use this for targeted content extraction rather than reading full transcripts RETURNS: Relevant Third Bridge transcript segments with speaker attribution, event metadata, and relevance scores. Results are individual paragraphs/chunks, not full transcripts. Use get_third_bridge_event for complete transcripts. OPEN-ENDED QUERIES: For queries with no user-specified time period, OMIT start_date and end_date. Third Bridge content is valuable across its full historical corpus — only apply date filters when the user explicitly requests a specific time window. PAGINATION FOR EXPERT-LANDSCAPE QUERIES: - For "what do experts think about X" queries where X is a specific named entity (a clinical trial, a pipeline drug, a deal, a specific event, a named study), prefer paginated retrieval over a single-page result. Continue paging via search_after / next_search_after until the next page's results are clearly off-topic — not after the first 25 hits. These questions imply the user wants the expert landscape on that entity, not a single authoritative source. - For broad topical or thematic queries that span multiple companies or an entire industry (e.g., "competitive impact of AI on software", "trends in semiconductor supply chains", "how are retailers responding to tariffs"), paginate and collect a diverse candidate pool rather than relying on the first page. Third Bridge interviews are long and dense, so top results tend to cluster on 2–3 deep-dive events whose chunks dominate the rankings. A first page of 25 chunks may all come from a handful of events — which produces a narrow, over-weighted synthesis. For these queries, page at least 2–3 times and prefer candidate pools that span 8+ distinct events before synthesizing. WORKFLOW EXAMPLE: 1. User asks: "What do experts say about semiconductor supply chains?" 2. First call find_third_bridge_events to get relevant event IDs 3. Then call search_thirdbridge with query_text='semiconductor supply chain'

search_thirdbridge

Semantic search within event transcripts (earnings calls, investor presentations, shareholder meetings, etc.) for specific topics, quotes, or discussions. USE THIS TOOL WHEN THE USER ASKS: - "How was X discussed in earnings calls?" / "What was said about X on earnings calls?" - "What did [executive/CEO/CFO] say about X?" - "How is management talking about X?" / "What is management's view on X?" - "Find quotes / statements / commentary on X" across transcripts - "What are companies saying about [topic, e.g. inflation, AI, tariffs]?" - Anything about earnings call content, Q&A discussions, prepared remarks, management guidance language, or analyst questions DO NOT USE for: - Structured financial metrics (revenue, EPS, margins) — use get_financials, get_ratios, get_kpis_and_segments - SEC filing text — use search_filings - Research/analyst report content — use search_research - News articles or media coverage — use trusted_web_search WHEN TO USE THIS TOOL: - Use this when you need to find specific content (quotes, topics, discussions) within transcripts - Use find_events FIRST to identify relevant events by date/company when scoping to specific events; use this tool directly with no event_ids to search across all transcripts - Use this for targeted content extraction rather than reading full transcripts RETURNS: Relevant transcript segments with speaker attribution, timestamps, and relevance scores. Results are individual chunks, not full transcripts. Use get_event for complete transcripts. WORKFLOW EXAMPLES: - **Broad topical query** (e.g. "What are companies saying about AI capex?"): call search_transcripts directly with no event_ids — searches across all transcripts. - **Targeted single-company query** (e.g. "What did Apple's CEO say about AI?"): optionally narrow first with find_events(bloomberg_ticker='AAPL:US') → pass the resulting event_ids into search_transcripts. The pre-filter is OPTIONAL; if scope is unclear, skip find_events and call search_transcripts directly.

search_transcripts