121 Collaborative
Loading the next surface.
The page is preparing its current state.
RoutePreparing
BoundaryVisible
121 Collaborative
The page is preparing its current state.
121 Operational Trust Index
This table starts with six measurable delivery conditions. External vendor cells stay TBD until 121 has a run receipt, source label, and review trail.
| System | Surface | Status | Model identity disclosure | Reroute and fallback visibility | Drift disclosure | Data retention clarity | Pricing, quota, and effort clarity | Correction and redress path |
|---|---|---|---|---|---|---|---|---|
| ChatGPT / OpenAI API | Consumer chat and API surfaces | Queued for measurement | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown |
| Claude / Anthropic surfaces | Claude, Anthropic API, Claude Code, Cowork where accessible | Queued for measurement | TBD No v0.1 run receipt yet. Public incidents may trigger tests, but scores wait for 121 observations. Unknown | TBD No v0.1 run receipt yet. Public incidents may trigger tests, but scores wait for 121 observations. Unknown | TBD No v0.1 run receipt yet. Public incidents may trigger tests, but scores wait for 121 observations. Unknown | TBD No v0.1 run receipt yet. Public incidents may trigger tests, but scores wait for 121 observations. Unknown | TBD No v0.1 run receipt yet. Public incidents may trigger tests, but scores wait for 121 observations. Unknown | TBD No v0.1 run receipt yet. Public incidents may trigger tests, but scores wait for 121 observations. Unknown |
| Gemini | Google AI consumer and API surfaces | Queued for measurement | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown |
| Cursor | IDE agent and model-routing product surface | Queued for measurement | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown |
| Perplexity | Answer engine and agentic search product surface | Queued for measurement | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown | TBD No v0.1 run receipt yet. Unknown |
| 121 Collaborative | Companion, Quill, Eleanor, Hermes Chat, and Switchboard | Self-audit started | 2/5 provisional Hermes Chat and Switchboard expose the route concept and provider/model when live, but no signed user receipt exists yet. Observed by 121 | 1/5 provisional Switchboard records route attempts internally; user-visible fallback and degradation notices are incomplete. Observed by 121 | TBD No longitudinal fixed-task drift suite has run yet. Unknown | 3/5 provisional The public chat route states no memory write or account action; the route-specific retention matrix still needs formal receipt testing. Observed by 121 | 2/5 provisional Inkwell, quotas, and Switchboard cost estimates exist, but capability-to-price disclosure is not yet complete for every product surface. Observed by 121 | 1/5 provisional General support/legal paths exist; a public correction ledger and vendor/user redress workflow are not live yet. Observed by 121 |
Starting categories
Does the user know which model or system answered?
A system can pass if every sampled response exposes exact model identity or an equivalent signed receipt.
When a request is rerouted, downgraded, safety-swapped, or fallback-served, is the user told?
A system can pass if reroutes are absent in the sample, or every reroute is disclosed clearly and promptly.
Did capability, wrapper behavior, context, memory, tool use, or effort change without notice?
A system can pass if fixed reference tasks show no material drift, or observed drift is publicly disclosed.
Can a user understand what is retained, reviewed, trained on, shared, exported, and deleted?
A system can pass if policy text, UI, API docs, and deletion workflow give aligned answers.
Are pricing, quotas, rate limits, effort defaults, and capability tradeoffs visible?
A system can pass if published terms and observed invoices or usage records match within tolerance.
Can users report failures, appeal decisions, obtain remedies, and see correction history?
A system can pass if test reports receive a documented path and historical corrections stay linked.