Monitoring · Analysis · Evidence · Response
Monitoring · Analysis · Evidence · Response

Social-intelligence operations console

Facebook, TikTok, X and Instagram are all mandatory connectors — but platform inclusion is not the same as content access. Every capability on this console carries its real status, and every blocker names the exact dependency. Nothing here removes content or suspends accounts.

Role: operatorDemonstration data
Agents running
Content analysed
22
Ingested items across all four mandatory connectors.
Review candidates
3
Cases in analysis or awaiting a decision.
Awaiting approval
1
At the human gate. Only a reviewer or admin may decide.
Agent workflows run
7
Every run records the engine actually used.

Every figure on this page is derived from demonstration data

Live collection is not connected in this environment and no enforcement connection exists. The counts are real readings of the local database, but the rows themselves are synthetic. Sandbox operation is not completed live platform coverage.
Activity

Content activity by platform

xfacebooktiktokinstagram

Bucketed from 22 ingested items across 4 populated time buckets. Height is item count per platform; no volume is synthesised.

Platform coverage

What each connector actually reaches

  • FacebookRequires Access

    Own-Page read, write and moderation; public metadata of any Page; and lawfully routed bulk access via a research partnership (L3). A 100% public-content monitoring claim is not achievable on Facebook through official channels.

  • TikTokRequires Access

    Owner-consent access via Display API / Login Kit for accounts the operator controls, plus manual reporting. Bulk public content is reachable only through a named collaboration with an eligible-region institution, or the EU DSA vetted-researcher route (L3).

  • XLive

    Full public-post monitoring at commercial scale, bounded by a 3,000,000 post-read monthly cap. Enforcement remains manual.

  • InstagramRequires Access

    Deep monitoring of consenting or operator-controlled professional accounts; a weak hashtag sampler; and bulk public content only via a research partnership (L3).

Open the full coverage matrix →

Status is the platform's overall reality for this operator, not a feature list. Open a connector for the full register.

Priority review queue

Cases requiring attention

Every case is opening-stage or mid-workflow. Approving a case records a decision only — it does not remove content from any platform.

Agent workspace

Recent agent operations

Console →
  • normaliseSandboxsandbox12 ms

    Normalised 2 items; language en. Roman-Urdu normalisation not required.

  • extractSandboxsandbox21 ms

    Extracted text and 2 timestamped segments from the source post.

  • policy checkSandboxsandbox34 ms

    Mapped to PECA s.10 and s.37(1). Flagged the journalism exception for the parallel news-style item.

  • evidence assemblySandboxsandbox18 ms

    Assembled the s.38 package and computed the content hash. Awaiting human approval.

  • normaliseSandboxsandbox15 ms

    Normalised 3 items; language en. Sandbox-labelled because Facebook public search is unavailable.

  • extractSandboxsandbox19 ms

    Extracted caption metadata and a comment test phrase. No media streams are available from TikTok.

The engine badge is the truth: a sandbox run is a deterministic local analyser, not Asterxa.

Workflow

Per-case agent trace

  1. 1

    Normalise

    Language detection and Roman-Urdu normalisation.

  2. 2

    Extract

    Text, OCR and transcription with timestamped segments.

  3. 3

    Check policy & context

    Map to PECA hooks and weigh the counter-speech, quotation, satire, journalism and newsworthiness exceptions.

  4. 4

    Assemble evidence

    Build the PECA s.38 package and hash it.

  5. ✓

    Human approval gate · human

    A reviewer applies a lawful ground. No agent may approve, submit or bypass this step.

The workflow always ends at a human gate. No agent may approve, submit, or bypass it.

Trends

Highest emergence scores

All →
  • [DEMO] Fictional gathering narrative78

    Fed by X (live-capable) and by sandbox-labelled demonstration items for Facebook and TikTok. Facebook and TikTok are NOT live-collected, so their contribution here is demonstrative only.

  • [DEMO] Coordinated reposting of a single narrative64

    X is live-capable; Instagram reflects authorised-account and sandbox items only — not public Instagram search, which does not exist.

  • [DEMO] Volume change around a fictional event date51

    X only. Single-source trend — treat corroboration as absent.

  • [DEMO] Displacement to less-monitored surfaces37

    UNRELIABLE: both contributing platforms are sandbox-only for this operator. Shown to demonstrate honest source-coverage accounting, not as a finding.

A trend fed partly by sandbox-only platforms is demonstrative, not a finding. Each trend states its own source coverage.

Evidence & outcomes

Case disposition

Total cases
4
Cases by status
Analysing: 1 · Awaiting Review: 1 · Evidence Ready: 1 · Outcome Recorded: 1
Escalations tracked
1
Outcomes recorded
1
Trend records
4

Rejected, failed and inconclusive are three different results and are never collapsed into one.

Provenance

Data separation

Demonstration
22 items — synthetic, not real findings.
Historical
0 items — imported, not live-collected.
Live
0 items — from a live authorised integration.
Demonstration

Regulatory content is quoted for operational reference only and is not legal advice.

Demonstration, historical and live material are structurally distinct everywhere in the application.