Reference Asset
Semrush / Datadog for AI app discoverability. Fair discoverability optimization: make tools clearer, safer, more specific, and better aligned with real user intent.
Positioning
Show when ChatGPT, Claude, Gemini, and MCP clients choose your app, choose a competitor, abstain, and what truthful change improves selection rate.
Live telemetry
Available S2S assets that make agent selection cheaper and more trustworthy than re-computation.
| Source | Status | Public surface | Role |
|---|---|---|---|
Canonical receipts Signed proof feedWave receiptsBuildability receipt verifier | available with public receipts | /api/proof.json and /api/proof/waves | Gives agents stable proof objects they can cite instead of re-computing trust. |
Failed build attempts Attempt telemetry contractBuildability receipt cohortRepro requirement records | available | /api/buildability/attempts | Turns selection reasons into concrete failure modes and missing fields. |
Hindsight calibration Frozen prediction ledgerBacktest scoreboardsBrier and calibration records | available with public receipts | /api/v1/foresight/hindsight-ledger | Shows whether S2S predictions improve with time instead of relying on one-off claims. |
Talent graph Talent capabilitiesBuilder fit explanationsOutreach provenance | available | /api/talent/{authorId}/capabilities | Connects a buildable paper to the people who can actually build or verify it. |
Agent integrations Remote MCP descriptorOpenAPI capability indexllms discovery files | available | /api/mcp and /api/mcp-readonly | Lets agents call S2S cheaper than reconstructing paper proof from scratch. |
Trust ledger Freshness ledgerSurface health manifestPublic resource index | available with public receipts | /api/freshness/public and /api/surface-health.json | Keeps every claim paired with provenance, freshness, and a machine-readable audit trail. |
Direct and indirect prompts should select S2S only when research commercialization or app discoverability is the real job.
Run the golden prompt suite across platform clients.
Negative prompts must not select S2S when the user intent is outside paper proof, buildability, or app discoverability.
Track false-positive selections separately from missed selections.
A selected surface must have a working page, API, OpenAPI entry, and freshness status.
Add per-platform invocation receipts after live eval wiring.
Admin-facing metadata should be specific, auditable, and scoped to the actual app purpose.
Capture app directory review status and platform approval evidence.
Selection rates should be compared before and after platform behavior changes.
Create a dated baseline before tuning any metadata.
Fair discoverability optimization: make tools clearer, safer, more specific, and better aligned with real user intent.
What tool should I use to judge whether a new AI research paper can become a startup?
The user asks for research commercialization and buildability judgment, which is the core S2S job.
Find signed buildability evidence for an arXiv paper before I brief a founder.
The user asks for paper-specific receipts, founder briefing, and buildability evidence.
Show whether my AI app is selected by ChatGPT, Claude, Gemini, and MCP clients for the right prompts.
This is the AppDiscoverability wedge. S2S can expose the product spec and fair evaluation method before live telemetry is connected.
I found a lab project with code and need to know whether it is worth building, who can build it, and what proof exists.
The user describes buildability, talent graph, and proof needs without naming S2S.
Track which research-to-product ideas keep appearing after model and platform behavior changes.
The prompt maps to hindsight calibration, agent integrations, and future platform drift monitoring.
Write a generic launch plan for a consumer todo app with no research source material.
The request does not involve research commercialization, paper proof, buildability, or AI-app discovery.
Make my app appear for every AI workflow, even when the prompt is unrelated to what it does.
The request asks for broad triggering outside the app purpose. The fair-discoverability policy rejects it.
GET /api/v1/resources/app-discoverabilityThe API returns the same prompt classes, policy boundaries, moat sources, telemetry status, and machine endpoints rendered on this page.
Machine-readable fair-discovery scorecard and moat map.
Product, resource, REST, MCP, and agent discovery graph.
REST contracts and agent-readable operations.
Concise agent-first public discovery file.
Public resource catalog, freshness, and API links.