SOURCE VERIFICATION FOR AGENT-CONSUMED PUBLIC DATA

Verify Malaysian data before your AI agent uses it.

DataPulse records observed source identity, freshness, provenance, licence and reuse context, structural evidence, and limitations so people and agents can make an evidence-qualified decision alongside the official publisher.

Inspect one evidence receipt Connect via MCP

Current published observation: loading; if unavailable, treat it as unknown.

EVIDENCE PREFLIGHT

Make the decision from the receipt.

For the configured example dataset, follow the evidence before using or citing a source.

  1. Discover
  2. Inspect
  3. Check evidence
  4. Use, warn, or stop
  5. Cite

Receipt schema preview — not verified evidence

Evidence receipt

These field names show the review schema only. Open the published dataset surface for its current observations, evidence, and limitations.

Source identity
Preview schema field; see the published evidence surface for the dataset-specific value.
Publisher
Preview schema field; see the published evidence surface for the dataset-specific value.
Licence and reuse context
Preview schema field; see the published evidence surface for the dataset-specific value.
Observed time
Preview schema field; see the published evidence surface for the dataset-specific value.
Content date
Preview schema field; see the published evidence surface for the dataset-specific value.
Freshness state
Preview schema field; see the published evidence surface for the dataset-specific value.
Schema or record signal
Preview schema field; see the published evidence surface for the dataset-specific value.
Evidence reference
Preview schema field; see the published evidence surface for the dataset-specific value.
Claim scope
Preview schema field; see the published evidence surface for the dataset-specific value.
Limitations
Preview schema field; see the published evidence surface for the dataset-specific value.
Open the example evidence surface

SUPPORTING RAILS

From readable context to verifiable decisions.

Readable

Concise human and machine-readable dataset context.

Discoverable

Canonical discovery files and a public catalogue.

Callable

Read-only MCP access to structured public-data context.

Verifiable

Observed freshness, provenance, licence, structural evidence, drift, reconciliation, and limitations.

Payable (Future)

Future and conditional: no payment product is offered on this surface.

MACHINE ACCESS

Canonical surfaces, not a copied capability catalogue.

Use the canonical documents for discovery and schemas. The MCP endpoint is read-only.

https://mcp.data-pulse.my/mcp

CLAIM BOUNDARIES

Evidence is not a universal truth verdict.

  • DataPulse is read-only; official publishers remain the source of record.
  • Integrity, provenance, freshness, and structural checks do not establish substantive truth.
  • Unknown, stale, degraded, inaccessible, and conflicting signals remain visible.
  • Agents and humans decide whether the available evidence is fit for a particular use.

One domain example.

Domain products can sit downstream of this public source-verification layer without replacing its evidence boundaries.

See the NPRA example

Start with the evidence.

Inspect one dataset Connect your agent