Arc Hub map

Network reachability, the open attention list, the next action and the most recent proof, each with the time it was observed. It reads live state and never caches a healthy answer.

The existing route derives actionable, caution and context states from the same bounded workspace facts.

The existing route reads the configured Arc endpoint and published mirrors. Metrics without a measured source stay unavailable.

Contract identity, interface, decimals and issuer come from the network registry and are confirmed against the chain. Transfer counts, volume and holder coverage need an indexer nobody serves for Arc yet, and render as unavailable rather than as a plausible number.

Every endpoint this deployment can reach is probed directly and reported with its own latency, chain match and error. Provider states that cannot be measured from here are named as unmeasured instead of assumed healthy.

There is no Flows route. H1 must index and normalize selected contracts, backfill a declared window and report partial coverage.

There is no Ecosystem route. H2 needs a verified project registry, named contracts and a measured activity method.

There is no Agent Economy route. H2 needs verified registry and job sources before Ryntra can show activity or reputation facts.

There is no markets route. It requires provider-backed executable quotes at standardized sizes and measured coverage; App Kit alone is not a market database.

The route prepares exact calldata, verifies an EIP-712 intent authorization, opens the connected wallet for the transaction, records the hash and reconciles the Arc result. The historical receipt proves only its own run.

The preview route exists. No swap has completed end to end, and the exact external-signing payload is not verified.

Three operations you can use now carry full weight; everything still waiting on a gate is grouped below them and says what it is waiting for.

A request encodes the exact amount, asset, recipient and purpose into a link and a QR code, and opens Send already filled in for whoever follows it. Nothing is created on a server and nobody is notified — tracking whether a request was paid needs a recipient and an asset on each recorded operation, and the ledger carries neither.

There is no Bridge route. One complete CCTP testnet lifecycle, including recovery, reconciliation and a receipt, must be proven first.

There is no Unified Balance route. Gateway integration, balance accounting, provider recovery and an independently reconciled run are required.

The preview route exists. Source contracts are implemented, but no live payout has been owner-signed and independently reconciled.

A CSV becomes a checked queue: address shape, duplicate detection, per-row limits and the batch total against the available balance. Each row is then authorized and signed one at a time through the Send lifecycle. A whole batch has not yet been executed end to end.

The payment-request lifecycle and its webhook delivery both have to exist first. Neither does.

A refund is only meaningful when it is bound to the original settled payment and cannot exceed it. That binding needs the checkout lifecycle it refunds, which does not exist yet.

Escrow holds someone else's money against a condition, which makes it a custody and dispute design before it is a screen. It stays registered as a direction until a design partner needs it.

Recurring value movement needs a delegated permission that can be revoked and proven revoked. Until the permission and revocation semantics are demonstrated end to end, a schedule would be a promise to move money that nobody can withdraw.

There is no FX route. StableFX eligibility, provider integration, legal boundaries and one complete settlement proof are required.

The existing route reads the active policy and its raw values one level deeper. It is not a configurable policy builder.

Labelled beneficiaries stay in this browser and fill Send and Treasury payout without retyping an address. They are a convenience, not an allowlist — the activation delay, approval and shared registry that make a recipient list a control do not exist yet.

Account, wallet, workspace presentation and privacy boundaries, with diagnostics collapsed. Ryntra holds no key and no seed phrase; the wallet section states exactly what it can and cannot revoke.

The existing Settings route shows the current connection and permission boundary. Ryntra stores no key and wallet permission is revoked inside the wallet.

The preview route exists. It describes designed roles and read-only tool names; no Ryntra agent authorizes, signs or executes.

Policies today are read as published rules. Authoring them — limits, reserves, allowed assets and recipients, fee ceilings, approver counts — needs a versioned editor with simulation, so a change can be understood before it blocks a payment.

There is no Approvers route. It needs tenant-scoped role assignments, activation history and receipt-bound approval evidence.

No Arc-native Agent Control route is enabled. The separate /agents/workspace route is a non-executing sample source preview; agent identity, policy enforcement, human authorization and recovery evidence must be proven before Arc controls can be enabled.

The existing route filters only the current session ledger and exposes execution, reconciliation, transaction and receipt state without inventing rows.

The existing route separates finalized session receipts from the one historical Arc reference proof and never manufactures PDF or share output.

A receipt is re-hashed from its own recorded content and compared with the hash it carries, so a modified receipt fails rather than fails to be noticed. It answers verified, not found, superseded or tampered, and needs no wallet.

Every operation this workspace recorded, read as a history: what was decided, what actually settled, where the two differed and what recovery was required. It is a projection of receipts, not a second record.

Counts, fees, outcomes and deviations summarised over the operations this workspace holds, with CSV and JSON export. Every figure is a sum of recorded operations; nothing is estimated or extrapolated.

The existing route documents source contracts over the same Evidence Kernel. It is a reference, not a public Arc Data API or key-issuance console.

Guard API contracts and examples exist in source. The public Arc data product, indexer projections, tenant gates, keys and quotas are not enabled.

The lifecycle exists in source and backs this workspace through a same-origin route bound to its own session. A partner-facing contract additionally needs tenant authentication, published versioning and a stability commitment, none of which is in place.

A source SDK and examples exist. Public package publication, versioned distribution and Arc data-client scope are not enabled.

There is no key console. Authentication, issuance, rotation, revocation, quotas, audit and recovery require their own security and release packet.

There is no public MCP surface. H2 needs the indexed data projections, tenant gates and tested read-only tools; signing and policy mutation stay prohibited.

There is no webhook product. Delivery contracts, authentication, retry, deduplication, audit and tenant controls must be implemented first.

Commercial accounting belongs to one shared Ryntra kernel, not a ledger per workspace. It waits on the developer platform having something metered to account for.