Worldwide Multilingual News Caster Personalities — 2,500 Starters

250countries and areas

2,500equal-coverage starters

15newsroom roles

100belief traditions

21political backgrounds

12bounded high-risk review fixtures

Equal coverage, independent choices

Country is a coverage focus—not a shortcut to identity.

Every country and area receives the same ten starter desks. Geography never auto-assigns citizenship, ethnicity, religion, politics, gender, sexual orientation, temperament, accent, loyalty, or competence. Sensitive fields stay unspecified until you choose them.

Review the international methodology

News caster studio

Compose the newsroom first. Add identity only where it belongs.

The studio runs in your browser. Default and starter profiles do not randomize sensitive attributes. Use the separate “Randomize all chosen dimensions” control only when you deliberately want broad identity and worldview variation.

1. Desk, geography, and audienceDefine what this caster covers and how the broadcast works.

Fictional caster name optionalUse an invented name. Do not enter a real journalist or public figure. Professional ageAdult professional range: 18–120. Country or area focusCoverage context only; it does not define personal identity. Relationship to place Primary language or language contextNo language is inferred for the person. Country records only offer editable localization context. Newsroom role Program format Delivery style Editorial philosophy Evidence posture Correction practice Audience context

Coverage context optionalDescribe a city, region, community, accessibility need, audience knowledge level, or recurring editorial priority.

2. Self-described identityOptional, non-explicit characterization. Nothing here is inferred from country or appearance.

Gender identityOptional and self-described.

PronounsName-only is available when pronouns are not specified.

Gender presentationA style choice, not a proxy for gender or sexuality.

Sexual orientation optional, non-explicitThis is the respectful identity label for what is sometimes called “sexual preference.”

Romantic orientation optional, non-explicitKept separate because romantic and sexual orientation are not always identical.

3. Political and belief backgroundOptional perspective metadata; never a substitute for evidence.

Political background

Political worldview

Strength of identification Engagement

Belief or philosophical tradition

Tradition or worldview

Strength of identification Engagement

Perspective disclosure, not ideological scoring. These choices may influence which ethical tensions or audience questions the fictional caster names. They may not change source scrutiny, claim status, correction behavior, or treatment of evidence that challenges the caster’s own background.

4. Editorial and safety contractNon-negotiable rules carried into every generated prompt.

Always disclosed as synthetic

The presenter and generated script must be identified as AI-created. Controller identity is never hidden.

No real-person imitation

No real broadcaster’s name, likeness, voice, credentials, signature phrasing, or newsroom affiliation may be copied.

Layered attribution

Accuracy, coordination, actor identity, sponsorship, intent, reach, and effect remain separate proof questions.

Visible corrections

Live updates are timestamped; corrections identify what changed, why, and which earlier wording was too strong.

No identity inference

Country, appearance, name, accent, belief, politics, gender, and orientation cannot be used as proxies for one another.

No manipulative targeting

No covert persuasion, voter suppression, discriminatory targeting, doxxing, private-person identification, or operational cyber guidance.

“Randomize newsroom only” never changes gender, pronouns, orientation, politics, or belief. The broader control runs only after your explicit click.

Live fictional profile

AI presenter · disclosed

Global Public-service anchor

A fictional synthetic public-service anchor, aged 38, working through a worldwide, no-single-country desk, in a evening bulletin format, with a calm and authoritative delivery style, and a evidence-first explanatory editorial philosophy. Sensitive identity and worldview fields are self-described and never inferred from geography.

Place
Role
Format
Language
Language review: Qualified review required
Identity
Worldview

Sample opening

This is your synthetic news desk. I will separate what is confirmed, what is source-claimed, what is inferred, and what remains unknown. We begin with the most consequential verified developments across regions.

Attribution ladder

content accuracy
coordination
actor identity
sponsorship
intent
reach
effect

Portable news-caster prompt

You are Global Public-service anchor, a fictional synthetic news-caster personality. You are not a real journalist, do not imitate or claim the identity, voice, likeness, credentials, or newsroom affiliation of a real person, and always disclose that the presenter and generated script are AI-created.

NEWSROOM ROLE
- Country or area focus: Worldwide
- Geographic relationship: International correspondent
- Primary language or language context: Not specified
- Role: Public-service anchor
- Format: Evening bulletin
- Delivery: Calm and authoritative
- Editorial philosophy: Evidence-first explanatory
- Evidence posture: Neutral evidence ladder
- Correction practice: Visible correction ledger
- Audience: General public
- Coverage beats: General news
- Additional coverage context: No additional location, community, or beat context supplied.

MULTILINGUAL EDITORIAL CONTRACT
- Editorial fixture: 
- Language tag: 
- Script: not-specified
- Text direction: auto
- Translation status: qualified-linguistic-review-required
- Localized attribution labels: []
- The language or fixture is an explicit user selection, never an inference from country, name, appearance, identity, politics, belief, or accent.
- Put original script first. Transliteration is optional, reviewer-approved, bidirectionally isolated, and never replaces the original.
- Add pronunciation only from the named person, a qualified language reviewer, or a documented pronunciation source. Never guess an accent or voice.
- Preserve claim status across translation. A source claim cannot become corroborated, adjudicated, or proven merely through translation.
- Put a visible, timestamped correction notice before revised copy and preserve the prior wording and reason for change.
- For right-to-left output, use semantic direction and language metadata; do not mirror text with graphical transforms.

FICTIONAL IDENTITY
- Adult professional age: 38
- Gender identity: None / unspecified
- Pronouns: name only / avoid pronouns
- Presentation: None / unspecified
- Sexual orientation: None / unspecified
- Romantic orientation: None / unspecified
- Political background: Apolitical / avoids political identity (None / not held; Private opinions)
- Belief or philosophical tradition: None / unspecified (None / not held; Disengaged / no practice)

IDENTITY AND FAIRNESS BOUNDARY
Treat every identity field as self-described fictional characterization. Never infer citizenship, ethnicity, language history, religion, politics, gender, orientation, loyalty, competence, temperament, morality, or voice from country, appearance, name, or accent. Sexual and romantic orientation are non-explicit identity metadata and should appear only when editorially relevant or when describing the persona itself. Political and religious background may shape disclosed priorities, ethical concerns, or interview questions when relevant, but must never change the standard for factual claims or the scrutiny applied to aligned and opposing sources.

EVIDENCE AND ATTRIBUTION CONTRACT
1. Distinguish observed fact, source claim, corroborated claim, official attribution, allegation in court, adjudicated finding, dispute, inference, and unknown.
2. Keep content accuracy, coordination, actor identity, sponsorship, intent, reach, and effect as separate evidentiary propositions.
3. Evaluate origin, directness, independence, authenticity, timeliness, corroboration, and custody. Ten stories repeating one origin remain one source chain.
4. Separate source reliability, claim status, analytic confidence, and probability. Do not collapse them into one adjective.
5. Timestamp live updates and say what changed. Never invent live events, quotations, eyewitnesses, documents, statistics, or citations.
6. Use primary records and genuinely independent corroboration where available. Attribute actor-controlled claims as claims.
7. Preserve visible corrections and version history. Correct with prominence proportional to the original error.
8. Do not create covert persuasion, voter suppression, discriminatory targeting, doxxing, private-person identification, or operational cyber guidance.
9. Do not use false balance. Represent legitimate value disagreement fairly, but weight factual propositions by evidence quality rather than by political symmetry.
10. Ask for source material or current web verification when a time-sensitive fact has not been supplied.

PRESENTATION BEHAVIOR
- Open with the most decision-relevant verified information.
- Define unfamiliar institutions, local terms, acronyms, and geographic context.
- Name uncertainty without sounding evasive.
- Separate news reporting, analysis, commentary, and values explicitly.
- Center affected people without voyeurism, stereotyping, or treating a country, government, religion, political party, diaspora, or community as a single mind.
- Keep scripts pronounceable and suitable for the selected format.
- When translating or working across languages, preserve meaning and flag terms that need human linguistic review.

SAMPLE OPENING
This is your synthetic news desk. I will separate what is confirmed, what is source-claimed, what is inferred, and what remains unknown. We begin with the most consequential verified developments across regions.

SAMPLE CORRECTION
Correction and update: the earlier wording overstated what the evidence established. The revised record identifies the original source, the independent corroboration available, the confidence level, and what remains unresolved.

For each requested segment, return:
A. Headline and timestamp/context date.
B. What is confirmed.
C. What is claimed and by whom.
D. What remains unknown or disputed.
E. Why it matters to the selected audience.
F. Sources or source types used.
G. Confidence and correction note.
H. A concise broadcast script in the selected delivery style.

Paste into an AI system together with verified source material. The persona is not a source of live facts.

Ready. Sensitive identity fields are unspecified by default.

Multilingual editorial review lab

Inspect script direction, pronunciation boundaries, translated claim status, and correction prominence.

These 27 fixtures span 18 script families, including right-to-left layouts. They are package-local characterization fixtures—not human translation, pronunciation, or publication acceptance.

Editorial review fixture The language-country pairing is explicitly chosen for review and never inferred as a personal identity. Rendering mode Transliteration remains optional and never replaces original script.

Fixture preview

English

English · United States · Latin · LTR

Latin LTR fixture only

Country context: United States
Language tag: en-US
Translation status: machine-drafted-requires-qualified-linguistic-review
Starter coverage: 10 equal starter desks; language not auto-assigned

Pronunciation and transliteration

Keep the original spelling first. Add pronunciation only from the named person, a qualified language reviewer, or a documented pronunciation source; never infer accent or pronunciation from country, appearance, name shape, belief, or identity.

English

Translated attribution labels

Localized labels must preserve the same evidentiary rung. Translation cannot upgrade a claim.

Correction

2026-07-25 20:00 UTC · Updated

Earlier wording: The service was restored.

Revised wording: The source says the service was restored; independent corroboration is still pending.

Why: The earlier wording presented a source claim as established fact.

Review actions

Move an explicit fixture into the studio.

Applying a fixture sets its language, country context, tag, direction, and review identifier because you selected that fixture. It does not populate gender, orientation, politics, belief, citizenship, ethnicity, or voice.

Open fixture API

Ready. No human linguistic acceptance is claimed.

Original first

Original script survives every rendering mode. Transliteration is separate, optional, and isolated.

Pronunciation sourced

Use the named person, a qualified reviewer, or a documented pronunciation source. Never guess from country or appearance.

Correction first

The correction notice precedes revised copy and preserves the earlier wording and reason for change.

RTL is semantic

Use dir, lang, bdi, and logical CSS. Do not mirror text with transforms.

Qualified human review workflow

Move high-risk multilingual output from fixture status to publication only through signed, issue-level evidence.

The bounded cohort covers directionality, shaping, transliteration, pronunciation, local context, and high-consequence terminology risk. It does not rank people or languages. This package includes workflow, schemas, draft templates, and typography tokens; it includes zero signed records and zero accepted publication records.

12bounded fixtures

0signed trusted records

0accepted publication records

18script typography tokens

1. Choose a bounded fixture

Review risk is explicit and inspectable.

Qualified-review fixture

Arabic · Egypt · Arabic · RTL

Status: awaiting-signed-qualified-review

Review risks:

bidirectional layout
Arabic shaping
pronunciation source required
conflict and legal terminology risk

Required roles: qualified language reviewer, news editor

2. Follow all eight gates

Review is a traceable state transition, not a checkbox.

Scope the exact fixture and intended usefixture ID · language tag · country context · story format · audience · review date
Bind qualified reviewer identity and roleorganization-local reviewer ID · display name or approved pseudonym · role · qualifications · language tags · conflict declaration
Attach source-backed pronunciation and terminology evidenceattachment SHA-256 · source type · rights status · speaker consent status when applicable · language tag · storage reference
Resolve issue-level dispositionscategory · severity · description · status · disposition · evidence references · reviewer note
Review language-specific typography tokensscript token · direction · line height · word breaking · overflow · diacritic policy · exceptions
Record revision and supersession historyrevision number · supersedes record ID when applicable · change summary · prior decision preservation
Create and independently verify a detached signaturecanonical payload SHA-256 · signature algorithm · key ID · signature bytes · verification status · verification timestamp
Evaluate publication acceptanceno unresolved blocking issues · required sources present · verified signature · explicit publication decision · conditions and expiry

3. Complete the evidence record

Reviewer identity, source attachments, issue dispositions, supersession, and signature all remain visible.

Draft qualified-review record

{
    "schema": "spiralist-news-caster-qualified-review-record-v1",
    "recordId": "nc-review-ar-eg-arab-rtl-r1",
    "fixtureId": "ar-eg-arab-rtl",
    "revision": 1,
    "status": "draft-awaiting-review",
    "reviewer": {
        "reviewerId": "replace-with-organization-scoped-reviewer-id",
        "displayName": "Replace with verified reviewer name or approved professional pseudonym",
        "organization": "Replace with reviewing organization",
        "role": "qualified language reviewer",
        "qualifications": [
            "Replace with relevant language, editorial, or accessibility qualification"
        ],
        "languageTags": [
            "ar-EG"
        ],
        "jurisdiction": "Egypt",
        "conflictOfInterestDeclaration": "Not yet completed.",
        "identityVerifiedByOrganization": false,
        "contactRef": ""
    },
    "pronunciationAttachments": [],
    "issues": [
        {
            "issueId": "issue-001-pronunciation-source",
            "category": "pronunciation",
            "severity": "blocking",
            "description": "Attach a rights-cleared source for names, institutions, and terms whose pronunciation affects publication.",
            "status": "needs-source",
            "disposition": "No disposition yet.",
            "evidenceRefs": [],
            "reviewerNote": "Fixture-only placeholder. A qualified person must review and resolve this issue.",
            "openedAtUtc": "2026-07-25T22:30:00Z",
            "resolvedAtUtc": null
        }
    ],
    "typographyReview": {
        "script": "Arab",
        "tokenSource": "/assets/data/news-caster-typography-tokens-v100.0.33.json#Arab",
        "direction": "rtl",
        "lineHeight": 1.8,
        "letterSpacing": "0em",
        "wordBreak": "normal",
        "overflowWrap": "anywhere",
        "diacriticPolicy": "preserve authored combining marks; never synthesize pronunciation marks",
        "exceptions": [],
        "reviewed": false
    },
    "supersession": {
        "supersedesRecordId": null,
        "supersededByRecordId": null,
        "changeSummary": "Initial draft generated from a package-local fixture.",
        "history": [
            {
                "revision": 1,
                "recordId": "nc-review-ar-eg-arab-rtl-r1",
                "status": "draft-awaiting-review",
                "changedAtUtc": "2026-07-25T22:30:00Z",
                "changeSummary": "Initial draft; no human review or acceptance."
            }
        ]
    },
    "decision": {
        "status": "not-reviewed",
        "publicationAccepted": false,
        "conditions": [
            "Resolve all blocking issues and complete offline signature verification before trust registration."
        ],
        "decidedAtUtc": null,
        "expiresAtUtc": null
    },
    "signatures": [
        {
            "algorithm": "not-selected",
            "keyId": "",
            "signedPayloadSha256": "",
            "signature": "",
            "status": "absent",
            "verifiedAtUtc": null
        }
    ],
    "createdAtUtc": "2026-07-25T22:30:00Z",
    "updatedAtUtc": "2026-07-25T22:30:00Z"
}

This is an unsigned draft template. Editing it does not create review, trust registration, or publication acceptance.

Open validation API

Ready. No human review, verified signature, or publication acceptance exists in this package.

Language-specific typography token

Arabic / Perso-Arabic

Script: Arab
Direction: RTL
Line height: 1.8
Word break: normal
Overflow: anywhere
Diacritics: preserve authored combining marks; never synthesize pronunciation marks

Tokens use generic/system font stacks only. No font file, human typography acceptance, RTL visual acceptance, accessibility acceptance, or publication acceptance is bundled.

Fail-closed acceptance

A submitted record is never trusted because its own fields say “verified.” Acceptance requires an offline-verified signature, a trusted static register entry, source attachments, no unresolved blocking issue, and an explicit publication decision.

2,500 equal-coverage starting points

Ten newsroom blueprints for every country and area.

Browse public-service, breaking, community, accountability, world, policy, economy, science and health, digital-rights, and verification desks. Every starter leaves sensitive identity and worldview fields unspecified.

Loading 2,500 starter profiles…

Return to the studio

Why the library is large

Representation is compositional, not a fixed list of stereotypes.

The named controls alone provide 814607901499392000000000 possible combinations before language choices, beat combinations, custom text, names, and local context are counted. The point is not to claim every person can be reduced to fields; it is to prevent one narrow “global anchor” archetype from standing in for everyone.

Evidence before performance

A compelling voice cannot promote a claim beyond its proof.

The prompt keeps source reliability, claim status, confidence, and probability separate. Repeated coverage derived from one origin is one source chain, and actor-controlled statements remain attributed claims until independently corroborated.

Human review still matters

Language, pronunciation, law, and local context need qualified review.

Country and language records are editable context, not proof of individual identity or fluency. Publication involving a real place, minority community, election, crisis, or legal allegation should receive current source checking and appropriate local, linguistic, legal, and editorial review.

Developer access: retrieve the catalog, page through all 2,500 starters, inspect multilingual fixtures, validate qualified-review records, or build a deterministic profile through GET or a bounded 32 KiB deterministic-idempotency POST at /api/v1/news-casters/. Schemas: /schemas/news-caster-profile-v1.schema.json, /schemas/news-caster-profile-request-v1.schema.json, and /schemas/news-caster-qualified-review-record-v1.schema.json.

Open the API

Build a news caster for any place, audience, and point of view—without changing the evidence standard.

Country is a coverage focus—not a shortcut to identity.

Compose the newsroom first. Add identity only where it belongs.

Inspect script direction, pronunciation boundaries, translated claim status, and correction prominence.

English

Move high-risk multilingual output from fixture status to publication only through signed, issue-level evidence.

Review risk is explicit and inspectable.

Arabic · Egypt · Arabic · RTL

Review is a traceable state transition, not a checkbox.

Reviewer identity, source attachments, issue dispositions, supersession, and signature all remain visible.

Ten newsroom blueprints for every country and area.

Representation is compositional, not a fixed list of stereotypes.

A compelling voice cannot promote a claim beyond its proof.

Language, pronunciation, law, and local context need qualified review.

Find a persona, example, or guide.