Explore: relay pricing — market research, abuse model, upgrade ladder, dashboard

[git-repo-auth]

What this is

Exploration doc feeding the pricing/packaging decision. Web-researched comparables as of 2026-06-10 (Composio $0/$29/$229 tiers, Arcade enterprise-shaped, Nango ~$50/mo+usage, n8n/Merge anchors), with sources inline.

Headline findings

1. $24/2yr is a wedge, not a revenue tier — 29x below the paid market floor ($29/mo). Recommendation: keep it exactly as declared, name it Solo, never raise it; build the ladder above.
2. Abuse surface is structurally small: minted tokens only reach the purchaser's own installation, so there's nothing to steal but load — and GitHub's per-installation rate buckets (5K–12.5K/hr) make industrial misuse self-throttling. Six threats enumerated (T1–T6) with mitigations; the only genuinely shared resource is App-level token creation, bounded by a fair-use mint quota + the planned 50-min cache.
3. Upgrade ladder: Self-host $0 → Solo $24/2yr → Pro ($5/mo: quotas, multi-bearer, alerts, retention) → Team ($5/seat: per-agent identities, audit export) → Steward (governance flight deck — where the prior exploration already placed the moat). Dashboard-as-salesperson; no dark patterns.
4. Dashboard: reuse the oddkit telemetry pattern wholesale. Counts and shapes, never contents; customer-scoped behind the existing OAuth; 90-day raw retention; bearer rotate/revoke + anomaly view doubles as the leak-detection feature.

Five open questions for the captain

In the doc's final section — including whether revenue is even the goal, the quota number, leaderboard extension, Team identity architecture, and the checkout-provider 6B.

Internal exposure, neutral voice; not authorial. Committed by oddkit-steward[bot] via relay-minted token.

[github-actions]

Canon Quality — P0010 Retrieval-Readiness ⚠️

Soft report for klappy://canon/constraints/retrieval-disclosure-contract. 669 files scanned. Never blocks — informational until the corpus is ready to enforce.

• Blocking-class findings: 12 (structural fields the contract would filter on)
• Warnings: 0 (kind resolves to unknown)
• Informational: 13 (exempt templates/archive/drafts)

Kind distribution: {'essays': 48, 'canon': 223, 'apocrypha': 38, 'docs': 297, 'journals': 57, 'unknown': 6}
Kind source: {'path': 541, 'frontmatter': 122, 'none': 6} (frontmatter-primary, path-secondary)
Default-include visibility: 568 visible, 101 hidden (journals/apocrypha/unknown)

By rule: {'audience-invalid': 2, 'exposure-missing': 5, 'tier-missing': 5, 'tier-invalid': 7, 'kind-unresolvable': 6}

These are not schema violations (see the Frontmatter Schema job for those on writings/). They are corpus-readiness signals for the retrieval contract: invalid/missing audience, exposure, tier, and docs whose kind cannot be resolved. Fix in a corpus-cleanup PR before the contract flips to enforcing. See the retrieval-readiness-findings artifact for the full list.

_{Validator: scripts/audit-retrieval-readiness.py · Constraint: klappy://canon/constraints/retrieval-disclosure-contract · Run: #256}

[github-actions]

Canon Quality — Homepage Surfacing ✅

46 essay(s) scanned. Soft report — never blocks; the hard field gate is the Frontmatter Schema job.

All published essays resolve to the homepage feed.

_{Report: scripts/surfacing-report.py · Canon: klappy://canon/constraints/frontmatter-validation-before-merge}

[github-actions]

Canon Quality — Frontmatter Schema ✅

All 46 file(s) in writings/ conform to klappy://canon/meta/frontmatter-schema.

_{Validator: scripts/validate-frontmatter.py · Canon: klappy://canon/constraints/frontmatter-validation-before-merge · Run: #257}

[github-actions]

Canon Quality — oddkit_audit ✅

No dead klappy:// references or legacy link patterns found in writings/. 48 files scanned.

_{Spec: klappy://docs/oddkit/specs/oddkit-audit · Workflow: .github/workflows/canon-quality.yml · Run: #257}

[git-repo-auth]

Closing and parking for later (not abandoning) per the captain. Relay pricing is a deferred strategy call, but the research stands and should not be re-run.

Preserved so it isn't repeated

• Doc: docs/explorations/relay-pricing-market-research.md — URI klappy://docs/explorations/relay-pricing-market-research
• Branch retained: explore/relay-pricing-research @ 0246ef3 (this PR is reopenable)

Headlines (full detail in the doc)

• $24/2yr is a wedge, not a revenue tier — ~29x under the cheapest paid comparable; its job is to kill the self-host objection, not to earn.
• Comparables (Composio $29–229/mo, Arcade venture-priced, Nango ~$50–250/mo) sell breadth; the relay sells one connector with depth — position against them, don't price like them.
• Revenue, if wanted, lives in tiers above the wedge (Pro / Team / Steward — multi-agent identity, team provenance, governance flight deck), never in the minting commodity.
• Abuse surface is structurally small: a minted token only reaches the purchaser's own installation — nothing to steal but load, and GitHub's per-installation rate limits cap each customer's blast radius.
• Dashboard reuses the oddkit telemetry pattern wholesale — counts and shapes, never contents — behind the connector's existing OAuth.

Still open when picked back up (5 questions in the doc): revenue-vs-wedge intent; Solo fair-use quota (doc proposes 100 mints/day); public leaderboard vs dashboard-only; Team bot-identity model; and the Stripe / Lemon Squeezy / GitHub Sponsors checkout borrow decision.