// live data for agents

the data your agent
can't find on Google.

Picabird turns an agent's question into clean, typed records from the source itself. Clinical trials, recalls, federal rules, energy regulation and SEC filings are live today. Product listings with live prices for shopping agents are in early access. Every record links to its source and says when we fetched it.

read the docs → tell us the dataset you can't get
output
live api

Every call above hits the real API at picabird.ai, which queries the source when you press run. The limit is 30 requests a minute.

how it works

A model turns your question into a query, and the response shows you that query. Code copies each record field from the source; no model writes record content.

   your agent                     picabird                                   official sources
 ┌──────────────┐   question   ┌──────────────────────────────┐   exact query   ┌───────────────────────┐
 │ "phase 2     │ ───────────▶ │ 1 interpret  (shown to you)  │ ──────────────▶ │ ClinicalTrials.gov    │
 │  oncology,   │              │ 2 fetch live (no stale index)│ ◀────────────── │ openFDA               │
 │  last 30d"   │ ◀─────────── │ 3 normalize  (typed fields)  │    raw JSON     │ Federal Register      │
 └──────────────┘   records    │ 4 cite       (url + time)    │                 │ SEC EDGAR             │
        ▲                      └──────────────┬───────────────┘                 │ FERC (via FR) · more ↓│
        │                                     │ watch (early access)             └───────────────────────┘
        └──────── webhook: record.created · record.updated {diff} ◀──┘

what agents use it for

This data sits in portals, PDFs, feeds and product pages that change daily. Live means you can run it in the console above. Early access means we're building it with design partners.

sources

Every record links back to the original filing, recall, registration or listing. "Refresh" is how often the source itself publishes.

sourcewhatrefresh (by source)status
ClinicalTrials.gov API v2trial registrations, status, endpoints, sponsorsdailylive
openFDAdrug, device and food recalls (enforcement reports)weeklylive
Federal Register APIproposed and final rules, notices, comment deadlineseach business daylive
SEC EDGAR full-text search8-K, 10-K and other filings by phraseas filedlive
FERC via Federal Registerinterconnection, transmission and tariff rules and noticeseach business daylive
CourtListener (Free Law Project)federal dockets and filingscontinuousearly access
Regulations.govrulemaking dockets and public commentscontinuousearly access
ISO/RTO interconnection queues, PUC docketsMISO, ERCOT, PJM, CAISO, SPP queues and state commission docketsvariesearly access
Marketplaces and brand storesevery listing in a category: comparable specs, the exact variant's live price, stock and landed costlive at queryearly access
County courts, council packets, customsstate and county filings, agendas, bills of ladingvariesearly access
// run the numbers

measured in your browser

Agent-native teams judge a data API on token efficiency, accuracy, completeness, speed, coverage and domain capability. Each card below made live calls from your browser when it scrolled into view. The accuracy and completeness cards also fetch the official source directly and compare it with our answer.

token efficiency

Typed records instead of raw source JSON.

·
waiting to run
accuracy

Every field cites the source path it came from.

·
waiting to run
completeness

Our total matches the source's own count.

·
waiting to run
speed

Live from the source, or warm from cache.

·
waiting to run
coverage

Which sources are live, and when each last updated.

·
waiting to run
domain capabilities

Domain terms become the source's exact filters.

·
waiting to run

Numbers are measured on this page load against today's data, so they move a little between runs. Rate-limited to 30 requests a minute; each card makes 1 to 4 calls.

built for the agent reading it

  • exact filters, shown back to youEvery response says how the question was read, e.g. "last 30 days" becomes updated_since 2026-10-05. An unsupported filter returns an error instead of being dropped silently.
  • notices for the model, warnings for youNotices are written for the model to act on. Warnings are written for the developer.
  • unknown stays nullA value the source didn't publish comes back null, never a guess. Summaries are assembled from fields by code.
  • watches with diffs early accessSave a query and get new and changed records POSTed to your webhook with a field-level diff, e.g. "primary endpoint changed".
  • agent-native surfacesAPI docs, /llms.txt, JSON or Accept: text/markdown, OpenAPI. An MCP server is coming.
// one record, trimmed (shape is picabird.record@1)
{
  "id": "ctgov:NCT07539311",
  "kind": "clinical_trial",
  "summary": "Phase 2 · Recruiting · Genmab ·
    advanced GI cancers · primary: ORR",
  "event": { "type": "updated", "date": "2026-11-03" },
  "fields": {
    "phase": ["PHASE2"], "status": "RECRUITING",
    "sponsor": "Genmab",
    "primary_outcomes": ["Confirmed ORR"]
  },
  "matched": [{ "criterion": "oncology",
    "evidence": "Advanced Gastrointestinal Cancers" }],
  "provenance": {
    "source": "ClinicalTrials.gov API v2",
    "url": "https://clinicaltrials.gov/study/NCT07539311",
    "retrieved_at": "2026-11-04T06:40:12Z",
    "source_updated_at": "2026-11-03T10:00:05",
    "paths": { "fields.phase":
      "protocolSection.designModule.phases" }
  }
}

tell us the dataset you can't get

We're early, and we're choosing what to build next with design partners. Tell us what your agent needs and how fresh it has to be. A person reads every request and replies.