Global investment research agent & knowledge base by ReportGem. Search 16 million+ global investment bank reports (Goldman Sachs, Citi, HSBC, UBS, JPMorgan, Morgan Stanley, Barclays, Nomura), 5 million+ reports from 200+ Chinese brokers and futures houses, company filings, SEC filings, earnings-call notes, academic papers and in-report charts; read page-cited source passages with sequential reading. OAuth login with a ReportGem account; new accounts get 10 free credits (about 100k returne
Scored across 5 tools
The four search_* tools map cleanly onto distinct corpora (filings, meetings, research, visuals), so an agent can pick by source type. The only soft boundary is between search_research (which can return real excerpts) and get_evidence (bounded page-cited excerpts / sequential reading), which an agent could plausibly reach for in the same situation.
Four tools use the identical search_<corpus> pattern and the fifth uses the standard get_<artifact> form, all snake_case verb_noun. The single verb change is semantically meaningful (retrieval vs. search), not an inconsistency.
Five tools is well within the ideal range for a search-and-retrieve server, and each one covers a different content type or operation with no filler. The surface is lean but complete for its scope.
The set covers the core lifecycle: find material across four source types and then pull verified, page-cited text from it. Minor gaps remain — no tool to browse coverage/source availability, no document-metadata lookup by ID, and no cross-source aggregation — but agents can work around these via the search tools.
Return bounded, page-cited source excerpts, or a sequential reading window (reading_mode="sequential" with next_cursor) for whole documents; unmatched material is never presented as verified evidence. 返回有界真实摘录或可续读的连续正文窗口；不能把无匹配的资料当作已验证证据。
Inputs: source_id (required), source_type (required), query, cursor, +3 more. Returns: object
Search US SEC filings and China listed-company announcements; returns verifiable internal source links. 检索美国 SEC 文件和中国公司公告，并返回可核验的内部来源链接。
Inputs: query (required), limit, market, date_to, +2 more. Returns: object
Search global and China earnings-call transcripts and meeting notes, distinguishing body-text evidence from listing-only hits. 检索全球和中国会议纪要、业绩会记录，区分正文证据与清单命中。
Inputs: query (required), limit, region, date_to, +2 more. Returns: object
Search sell-side research reports and academic papers by company, ticker, broker, industry, theme or date; returns bounded metadata plus real excerpts when the indexed body matches. 检索研报和论文；返回有界元数据，并在已索引正文命中时返回实际摘录。
Inputs: query (required), limit, date_to, sources, +1 more. Returns: object
Search indexed charts, tables, industry diagrams and research-relevant images with report/page provenance. 检索研报中的图表、表格、行业示意图，返回所属研报与页码。
Inputs: query (required), limit, date_to, date_from, +1 more. Returns: object
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Research layer for AI agents: stock theses, filing Q&A with page cites, monitoring. Research-only.
Grounded financial research for agents: filings, transcripts, news across US and Asia markets.
Investment research superagent: podcasts, SEC filings, and no-code research pipelines.
AI research on U.S. stocks, ETFs and crypto. Read your research reports and the method behind each analysis, ask SAGE about them, and price and start new research. Figures are computed in code on licensed market data, with sources listed. Research, not investment advice.
Add one secure layer between your agents and this server.
