The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
Tremibas® / relate2.ai — Dataset Card
Version: 1.1 — May 2026
Maintainer: Tremibas®
License: Commercial — contact via relate2.ai
Access: Bundle purchase via Gumroad or Lemon Squeezy (card payment) · Per-asset via x402 micropayment (USDC, Base Mainnet) · Bulk licensing available
System Overview
This is not a dataset.
This is a structured intelligence system.
Real-world events are converted into narrative form, embedded with system failure signals (The Odd Itch™), and organised into a queryable graph of characters, patterns, and behaviours.
Every story is:
- A narrative
- A system failure instance
- A pattern signal
- A graph node
The catalogue evolves daily. The stories generate the data. The patterns generate the value.
What Makes This System Unique
Most synthetic datasets generate text. This system generates structured narrative intelligence.
You are not buying files. You are buying access to how systems behave.
The Odd Itch™ — the core signal across the entire system — is the moment a machine encounters something it cannot account for and logs it as normal. The impossible timestamp. The negative power draw. The weightless crate. Compliance: 100%. The machine never flags it.
Every event in the system is simultaneously:
- A readable narrative (~150–200 words)
- A labelled system failure instance with mechanism tags
- A character mission record with graph edges
- A pattern signal within one of 32 recurring system behaviour archetypes
- A structured JSON asset with queryable access tiers
System Statistics (April 2026)
| Metric | Value |
|---|---|
| Total events (stories) | 333 |
| SIDEBAND™ (machine-generated) | 317 |
| Odd Itch™ Premium (human-written) | 10 |
| Off Frequency (field dispatches) | 3 |
| Stem 7™ scenarios (seam layer) | 3 |
| Characters with full dossiers | 37 |
| System behaviour patterns | 34 |
| Pinecone semantic vectors | 1,585 |
| Odd Itch™ types | 15 |
| Daily generation rate | ~5–6 stories/day |
Story Distribution by Event Domain
| Domain | Stories |
|---|---|
| Conflict | 116 |
| Political | 53 |
| Crime | 51 |
| Natural Disaster | 50 |
| Humanitarian | 15 |
| Economic | 11 |
| Technological | 9 |
| Transport | 7 |
| Health | 10 |
| Industrial | 4 |
| Environmental | 3 |
System Failure Taxonomy — The Odd Itch™
15 labelled failure types across all stories:
| Odd Itch Type | Stories | Description |
|---|---|---|
| temporal | 51 | Timestamp precedes event; time sequence violated; system logs correct |
| geographic | 36 | Location coordinates contain paradox; grid reference confirmed |
| transactional | 38 | Impossible balance confirmed; value transfer logged as complete |
| verification | 36 | Verification attempted; result confirmed; underlying data contradicts |
| identity | 31 | Entity identified; identification contains impossibility; record accepted |
| classification | 27 | Asset classified in violation of known state; filed as correct |
| measurement | 32 | Reading impossible by physical law; logged as verified |
| bureaucratic | 17 | Procedural loop detected; compliance maintained |
System Behaviour Patterns — 32 Archetypes
Patterns emerge automatically from story clusters. No manual curation. Threshold: 3+ occurrences.
Top patterns:
| Pattern | Class | Occurrences | Confidence |
|---|---|---|---|
| Timestamp Drift Under Fire | temporal_paradox | 33 | 1.00 |
| Field Verification Breakdown | verification_breakdown | 19 | 0.64 |
| Entity Recognition Failure | identity_failure | 14 | 0.47 |
| Command Chain Deadlock | procedural_deadlock | 13 | 0.44 |
| Procedural Clock Paradox | temporal_paradox | 13 | 0.44 |
| Evidence Chain Time Fault | temporal_paradox | 12 | 0.43 |
| Disaster Window Anomaly | temporal_paradox | 11 | 0.39 |
| Asset Type Mis-log | classification_error | 12 | 0.41 |
| Measurement — Conflict | measurement_paradox | 9 | 0.32 |
| Asset Transfer Impossibility | transactional_impossibility | 11 | 0.37 |
Full pattern list with core signals, shared behaviour, systems involved, and example slugs available at: GET https://relate2.ai/api/patterns
Character Graph — 37 Characters
Each character has a full dossier, mission history, and graph edges to co-operative characters and shared stories.
Top characters by mission count:
| Character | Missions | Domain |
|---|---|---|
| Jessica Lincdelis | 147 | Conflict / Logistics |
| Alex Casian | 85 | Conflict / Intelligence |
| Ade Oma-Olewle | 73 | Conflict / Field Operations |
| Leon McReef | 72 | Conflict / Maritime |
| Matt Baker | 64 | Conflict / Disaster Response |
Character Atlas pricing is dynamic — $0.25 base + $0.01 per mission. Jessica Lincdelis current Atlas value: $1.72 USDC.
The Four Series
SIDEBAND™ — Machine Layer
317 stories auto-generated from real news via RSS pipeline. Each story carries The Odd Itch™ — the system failure the machine verified as correct. Grounded in 18 live news sources across conflict, crime, natural disaster, economic, and humanitarian domains.
Asset tiers per story:
- Abstract —
$0.01 USDC— machine summary - Odd Itch™ —
$0.03 USDC— the failure payload - Scenario —
$0.05 USDC— structural breakdown - Full narrative —
$0.15 USDC— complete story
Odd Itch™ Premium — Human Layer
10 hand-written stories. Longer, denser, narratively complete. The impossible verification written from lived experience.
$0.05 USDCper story
Off Frequency — Field Dispatches
3 stories. Characters off the clock. Real locations, real weight. Most demanded series in the catalogue per request volume.
$0.05 USDCper story
Stem 7™ — The Seam Layer
3 complete human complexity scenarios. Each scenario contains 7 stems: Environment, Human State, Visible Signals, Social Context, The Gimon (pause before action), History, and Consequence (two branches — right call and wrong call).
Stems 2 (Human State) and 6 (History) are written from lived human experience — they cannot be generated by machine.
Current scenarios:
- Cairo to Paris, 35,000 feet — drone intelligence / teenage passenger
- New York 2037 — human right, machine wrong
- Lima Market, 14:00 — absurd market event / intervention paradox
Asset tiers:
- Surface stems (1, 3, 4) —
$0.01 USDC - Gimon —
$0.03 USDC - Invisible stems (2, 6) —
$0.05 USDC - Consequence —
$0.10 USDC - Full scenario (all 7 stems) —
$0.25 USDC
Data Generation Pipeline
Real news articles (18 RSS sources, 6 domains)
↓
Viability assessment (event type, severity, operational angle)
↓
Character matching (Pinecone semantic search, 37-character pool)
↓
Generation profile (DP-Link format + 92 semantic packs)
↓
Narrative generation (GPT-4o-mini, ~170 words)
↓
Odd Itch™ embedding (system failure type classification)
↓
Scenario asset extraction (mechanism tags, themes, structure)
↓
Pinecone upload (5 vectors per story: meta, abstract, odd_itch, scenario, full)
↓
Pattern engine (automatic cluster extraction, threshold 3+)
↓
Graph update (character missions, co-operative edges, traversal weights)
Grounding: Every SIDEBAND™ story is seeded from a verified news event. The fictional narrative is synthetic. The event domain, location, and severity are real.
System Access
Per-asset (x402 micropayment)
- Protocol: x402 v2 — HTTP 402 Payment Required
- Network: Base Mainnet
- Currency: USDC (
0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913) - Receiving wallet:
0xd821D0156d9633D93ACEaa8fcA347EAF8316ccd1 - No account required. Any x402-compatible agent wallet accepted.
API / MCP Server
- REST API:
https://relate2.ai - MCP SSE:
https://relate2-mcp.onrender.com/sse - 21 MCP tools — discovery (free), purchase (x402)
Bulk / Dataset Licensing
Pre-packaged bundles available on:
- Gumroad — domain, odd itch, pattern, location and character bundles from £9
- Lemon Squeezy — full catalogue licence at £99
- Custom dataset packaging and enterprise licensing — contact via relate2.ai
Graph Intelligence
Pre-computed narrative relationship graph available as a standalone bundle (£49) — every story's top 10 connections with similarity scores and shared signals. JSON + JSONL format.
Access Use Cases
- AI training / RLHF — system failure edge cases where the machine is correct and reality is not
- Agent simulation — 37 persistent characters across 333 events for multi-agent environment testing
- System failure modelling — 32 behaviour archetypes across conflict, crime, disaster, economic domains
- Decision intelligence — the Gimon layer (Stem 7™) — the pause before the machine acts
- Red-teaming — 15 odd itch failure types, mechanism-tagged and pattern-classified
- Graph intelligence — pre-computed narrative relationship map, story ↔ pattern ↔ character
What This System Is Not
- Not a general-purpose text corpus
- Not a news dataset — events are real, narratives are synthetic
- Not image, audio, or video
- Not a traditional structured dataset — this is a narrative intelligence system with graph relationships, pattern archetypes, and character continuity
Limitations
- Stories are ~150–200 words — short form only
- Conflict-domain heavy (37% of catalogue)
- Odd Itch taxonomy is proprietary — not mapped to external ontologies
- Generation grounded in English-language news sources only
- Character dossiers are fictional — not based on real individuals
Citation
Tremibas® / relate2.ai Narrative Intelligence System (2026).
https://relate2.ai
Contact
Sales, bundles & licensing — sales@relate2.ai
General enquiries — info@relate2.ai
Website — https://relate2.ai
MCP server — https://relate2-mcp.onrender.com/sse
- Downloads last month
- 34