OMS and inventory basics
What an OMS does, what inventory truth means, and why the customer promise breaks across channels.
Read basicsResources
Learn the operating model before you commit.
Commerce explainers, developer references, trust notes, AI-readable files, and preview-status pages. Use this as the map from operating problem to KarmanFlow proof.
ResourcesProblem to proof
LearnOMS, inventory truth, governed actionsBuildGraphQL, CLI, MCP, connectorsProvereceipts, security, preview statusMeasurepage views, docs search, attributionAgent-readllms.txt and full reference bundle
Start here
Understand the commerce problem first.
KarmanFlow is easiest to evaluate when the language is clear: promise, action, policy, receipt, and evidence.
Traditional stack vs KarmanFlow
A clear comparison of scattered handoffs versus one governed request, policy, receipt, and event path.
Compare modelsPreview, proof, and pending evidence
The honest status board for availability, ingestion, connector publication, agents, auditability, and scale evidence.
See statusCommerce lifecycle
Where KarmanFlow fits before the buy, after the buy, and across exceptions, returns, and support asks.
Walk lifecycleDeveloper knowledge
Build against the same boundary operators use.
References are public so builders can inspect the contract even while hosted tenant access remains preview-gated.
GraphQL guide
Typed product, inventory, order, receipt, and webhook surfaces for preview builders.
Open GraphQLCLI guide
Inspection-first workflows for tenants, inventory, events, receipts, connectors, and ingestion runs.
Open CLIMCP and agents
How agent tools should stay scoped, policy-aware, receipt-backed, and approval-bound.
Open MCPWebhooks and connectors
Replayable delivery, preview connector packages, source authority, and target-system publication posture.
Open connectorsProof and trust
Know what is preview, what is implemented, and what still needs evidence.
These pages keep availability, security, privacy, connector, and roadmap language tied to what the repo can prove.
Trust
Audit model, tenant isolation posture, receipts, policy decisions, and evidence still pending for production claims.
Open trustSecurity
Security posture, no hidden write paths, connector boundaries, and the public-site privacy model.
Open securityChangelog
What changed on the public site and docs surface, with the product status language kept honest.
Open changelogLabs
Preview capabilities, in-design work, and proof still needed before stronger availability claims.
Open labsAgent-readable
Resources for AI crawlers and internal assistants.
The public docs expose compact and full-text bundles so agents can cite source pages instead of guessing.
llms.txt
A compact map of the public KarmanFlow site for AI crawlers and agent summarizers.
Open llms.txtllms-full.txt
Plain-text export of generated public developer reference pages for agent retrieval and review.
Open full bundleReference search
Searchable static docs index for GraphQL, CLI, receipts, ingestion, and connector reference pages.
Search referencePreview
Turn the resources into a proof path.
Bring one workflow: availability, ingestion, connector replay, support context, or agent approvals. We will map it to the current preview surface and call out what is still pending.