Developers and agents

Give your coding agent the Bumpgrade public contracts first.

Bumpgrade publishes source-data routes, an agent manifest, public docs, llms.txt, and write-safety rules so agents can cite evidence, follow stable IDs, and respect permission boundaries before they suggest changes.

Start here

What agents can read and where to begin

Public resources are safe to read. Writes remain owner-gated until a route names the confirmation, idempotency, stale-state, audit, and redaction rules.

Read contracts

Core public contracts

The full manifest lists 23 public contracts. Start with these high-signal routes, then use the manifest when an agent needs exact stable IDs or boundaries.

Surface overview
PublicOpen

Feature catalog

Read feature status, Distinguish live from pending, Cite feature evidence

Stable IDsfeatureId, issue, status
BoundaryFeature status changes must land through GitHub issue/PR work and admin work-log updates.
PublicOpen

Public roadmap

Read public-safe roadmap state, Find blockers and next milestones, Cite issue evidence

Stable IDsroadmapItemId, featureId, issue, status
BoundaryRoadmap moves require an issue/PR or approved admin append path, not chat-only edits.
PublicOpen

Competitor comparisons and source evidence

Resolve competitor claims, Read retrieved dates, Cite official source URLs

Stable IDscompetitorId, sourceId, seoTargetId
BoundaryRefresh competitor claims from official sources before changing dated pricing, packaging, or feature claims.
PublicOpen

Commerce contract

Read redacted commerce architecture, Separate sandbox from live billing, Inspect referral attribution evidence, Inspect review-only commission ledger evidence, Inspect owner review action boundaries, Inspect non-billing post-purchase decision evidence, Inspect write safety rules

Stable IDsproductId, priceId, checkoutIntentId, referralClickId, referralAttributionId, reviewOnlyCommissionLedgerId, commissionReviewActionId, postPurchaseDecisionId, auditCorrelationId
BoundaryNon-billing post-purchase decisions can be recorded only for trusted checkout state; billing-impacting and payable commission writes require exact confirmation, idempotency, stale-state checks, audit correlation, owner review, and webhook evidence.
PublicOpen

Admin source-data bundle

Read public-safe work-log entries, Read user journeys, Read Mark attention summaries

Stable IDsworkLogEntryId, userJourneyId, markAttentionId, roadmapItemId
BoundaryHuman admin pages require Better Auth; agent writes need approved scripts or future confirmed APIs.
PublicOpen

Agent manifest

Discover read contracts, Route to source-data APIs, Understand write boundaries

Stable IDsreadContractId, mcpPlanId, agentDocId
BoundaryThis route is read-only until confirmed-write agent APIs exist.

Docs

Agent documentation map

liveOpen

Agent docs index

Human and agent-friendly index for Bumpgrade read contracts, source evidence, MCP direction, and safety boundaries.

liveOpen

Bumpgrade agent surface

Orientation for what agents can read, what is planned, and what requires owner credentials.

liveOpen

Bumpgrade commerce contract

Stripe sandbox, checkout, webhook, billing, and confirmed-write safety boundaries.

liveOpen

Bumpgrade source evidence

How public claims resolve to source IDs, URLs, issues, PRs, work-log entries, and caveats.

liveOpen

Bumpgrade admin surfaces

Which admin pages require owner auth and which source-data routes are public-safe for agents.

liveOpen

Bumpgrade MCP roadmap

First MCP resources and tools planned on top of the same public-safe contracts.

MCP and writes

Tooling builds on the same contract, with stronger safety.

MCP roadmap

Ready contracts

32 MCP resource contracts are backed by live JSON routes.

Confirmed writes

Public, billing-impacting, admin, and creator-speech writes need confirmation, idempotency, stale-state checks, audit correlation, and redaction.

Source grounding

Public claims need source IDs, issue or PR evidence, work-log records, source URLs, and retrieved dates.