Agentii Data Layer

Context Layer
for agent-use-ready data

Agentii turns filings, trials, approvals, events, and market signals into structured context agents can rank, cite, and act on without reading every raw document first.

Get API KeyAPI ReferenceView Coverage
SEC Filings1.2M pages
Clinical Trials48K studies
FDA Approvals12.8K
FAERS892K events
Earnings5.2K events
Profiles506 names
SCROLL
Source to context pipeline

What Data We Serve

Each source is normalized into the same agent-use-ready contract: description, keywords, metadata, page_content, and data. The Context Layer helps agents triage first, then retrieve exact cited evidence only when needed.

01Ingest
02Normalize
03Label
04Retrieve
SEC Filings
1.2M pages
Live
  • 10-K, 10-Q, 8-K filings
  • Earnings transcripts and presentations
  • Page-level retrieval with cited markdown
Endpoints:search_documents · get_document_pages · get_document_outline
Clinical Trials
48,291 studies
Q3
  • ClinicalTrials.gov records
  • Phase, sponsor, drug, condition, status
  • Signal labels for fast biotech triage
Endpoints:search_clinical_trials · get_clinical_trial
FDA Approvals
12,847 applications
Q3
  • NDA, BLA, ANDA approval records
  • Priority review and breakthrough status
  • Regulatory signal labels for agents
Endpoints:search_fda_approvals · get_fda_approval
FAERS Adverse Events
892,441 reports
Q3
  • Drug-reaction pairs
  • Seriousness and reporter classifications
  • Safety signal context for surveillance
Endpoints:search_faers_events · get_faers_event
Agentic Retrieval
Unified across all sources
Live
  • 4-step pattern: list, outline, search, read
  • Description and keyword triage before full reads
  • Row-level page_content access for citations
Endpoints:list_sources · read_source_outline · search_keyword_in_source · read_source_pages
Context Layer outputdescriptionkeywordsmetadatapage_contentdata

Agent-Use-Ready Format

Every record returns five fields optimized for LLM consumption. Agents triage with description + keywords, then fetch page_content only when needed.

metadata
Flat key-value pairs: ticker, date, phase, nct_id. Structured for filtering.
description
One-line summary for fast triage. Agents scan 100 descriptions in ~2K tokens.
keywords
Up to 10 high-signal terms. Enables keyword matching without reading content.
page_content
Full markdown rendering. The text agents feed to LLMs for deep reasoning.
data
Complete silver record with all columns. For programmatic access.
Ready for agents

Start querying the context layer

Give your AI agent a financial data layer it can actually use: source-normalized records, cited page content, metadata for filtering, and retrieval patterns designed for production workflows.

Get API KeyRead Docs
48agent-ready endpoints
10M+citation-backed pages
5,000requests per second
1,500free credits per month